https://orcid.org/0000-0002-7989-1031
Figures
Abstract
Lyme disease is a vector-borne illness transmitted by infected Ixodes spp. ticks. Dissemination of the Lyme spirochete—Borrelia burgdorferi—from the tick bite site results in a bi-phasic infection; the latter phase can cause severe musculoskeletal disease including arthritis. Lyme arthritis is an inflammatory disorder and maladaptive immune response to B. burgdorferi infection and its cellular products. One such product, which has been implicated as a key mediator of Lyme arthritis, is peptidoglycan. Peptidoglycan (PG) is a near ubiquitous feature of the bacterial cell envelope, but several chemical features make B. burgdorferi PG distinct from other members of the kingdom. We hypothesized the overall chemical composition and structural architecture of the B. burgdorferi cell wall are essential to Lyme disease pathogenesis. To manipulate the PG peptide chemical composition, as well as the native cross-links, we produced an isogenic deletion of a putative PG carboxypeptidase dacA homologue and assessed both the molecular and cellular phenotypes while probing the pathogenicity of our mutant strain. Our combined and comprehensive approach indicates while changes to PG stem peptide and cross-linking have virtually no discernable impact on any B. burgdorferi characteristic in vitro, alterations have significant impacts on tissue tropism and result in a near complete attenuation of Lyme arthritis. PG sacculi containing increased amounts of free and cross-linked pentapeptide surprisingly caused the disassociation of p83/100, an abundant periplasmic protein of unknown function previously implicated in joint tropism, likely contributing to a marked decrease in pathogenicity. These studies strengthen our understanding of the B. burgdorferi cell envelope, its unusual components, and further define bacterial features that mediate infectious arthritis.
Author summary
Many of the pathogenic mechanisms by which Borrelia burgdorferi causes disease involve cell envelope interactions with host components. To further elucidate the role of peptidoglycan in Lyme disease pathology we assessed a mutant strain with altered proportions of glycopeptides and found a near complete attenuation of arthritis. Cellular and molecular studies indicate this phenotype could be partially explained by disrupting the interactions between peptidoglycan and a previously uncharacterized periplasmic protein known to impact tissue tropism. Our findings suggest that the complex cell wall of B. burgdorferi, and its associated factors, are important contributors to Lyme disease pathogenesis.
Citation: Ahmad SS, Ebohon O, McClune ME, Trimble RN, Kellogg CN, Booth CJ, et al. (2026) Peptidoglycan architecture dictates protein interactions, tissue tropism, and arthritis in the Lyme disease spirochete Borrelia burgdorferi. PLoS Pathog 22(1): e1013849. https://doi.org/10.1371/journal.ppat.1013849
Editor: Sébastien Bontemps-Gallo, Centre National de la Recherche Scientifique, FRANCE
Received: August 18, 2025; Accepted: December 26, 2025; Published: January 20, 2026
Copyright: © 2026 Ahmad et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are in the manuscript and its Supporting information files.
Funding: National Institute of Allergy and Infectious Diseases (R01AI173256, R01AI178711, P01AI181934 to BLJ; R21AI173835 to WRZ) and USDA (VA-160113 to BLJ). SSA was partially supported by the Global Lyme Alliance, Bay Area Lyme Foundation, and the Steven & Alexandra Cohen Foundation awarded to BLJ. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Lyme disease (or Lyme Borreliosis) is one of the most reported bacterial infections in the United States [1]. Several biotic and abiotic factors have driven the expansive increase in both case numbers and geographical region of the tick-borne illness [2]. While the increased prevalence in recent years has garnered considerable attention from physicians, patients, and advocates alike, genetic evidence from historic remains coupled with phylogenetic reconstructions collectively suggest that humans have dealt with the consequences of Lyme disease for centuries [3,4]. All past, present, and future projections predict Lyme disease will continue to plague human health.
Clinical manifestations of Lyme disease are diverse and multifactorial. The acute stage of disease occurs within days of spirochete transmission from an infected Ixodes spp. tick [5]. Apart from skin rash(es) that may occur in some patients, initial symptoms are often vague and mimic many other illnesses. Without proper intervention, a disseminated infection will occur within weeks to months [6]. These later stages of disease can cause heart and neurological complications, but the most common illness in the United States is arthritis [5,7]. Inherent host factors, along with bacteriological products that may influence pathogenicity and/or tissue tropism, are thought to predicate both the disease presentation and severity [8].
Many of the typical virulence factors that pathogens produce to cause disease are missing from the Borrelia burgdorferi genome, the primary agent of Lyme disease in North America. For instance, classic effectors, and the machinery necessary to deploy them, are not thought to be produced by B. burgdorferi [9]. Both specialized toxins and those considered more broadly conserved in diderms (e.g., lipopolysaccharide/endotoxin) are also conspicuously absent from Borrelia species and strains known to cause human Lyme disease [10–12]. Thus, it is likely that basic cellular products associated with the unique bacterium’s tick vector-vertebrate host life cycle also act as virulence factors in susceptible hosts and tip the balance towards a particular clinical presentation.
The peptidoglycan (PG) of B. burgdorferi has recently been implicated as one such factor which could potentiate a robust and maladaptive immune response in certain hosts [13–15]. PG is a hallmark of virtually all bacterial cell envelopes where it acts as both a loadbearing structure and osmoprotectant, but it has also been implicated in pathogenesis where it may act as a virulence factor by modulating host responses in several instances [16–18]. Not only is B. burgdorferi PG released during growth where it may interact with host immune cells, but its chemical properties are unique [13]. For example, unlike most bacteria, the diamine involved in cross-linking peptides to adjacent glycan strands is L-Ornithine [13,19,20]. Alterations to the aforementioned glycan strands are also apparent since Borrelia spp. produce PG with a disaccharide of N-acetyl-glucosamine which is thought to contribute to the exceptionally long half-life of B. burgdorferi PG in mouse models of Lyme disease infection [14,20]. Lessons from mouse studies may extend to human Lyme arthritis since patients possess detectable levels of the cell wall biopolymer in the synovial fluid of their swollen joints [13,14]. How these peculiarities impact the overall severity and/or outcome of a B. burgdorferi infection remains poorly understood.
We postulated that alterations to the overall chemical architecture of B. burgdorferi PG may impact the host response and thus pathogenesis. PG is, however, essential and gross perturbations to salient features in the B. burgdorferi cell wall are unlikely to be tolerated. Here, we describe and characterize a mutant strain of B. burgdorferi with altered cell wall properties where the nascent sacculus possesses exceptionally high levels of non-cross-linked pentapeptide. Despite behaving nearly identical to a parental control strain in vitro, in vivo studies using a mouse model of infection suggest a near complete attenuation of Lyme arthritis. To better understand this response, we analyzed the B. burgdorferi cell envelope of our mutant and discovered that p83/100—an abundant periplasmic protein implicated in joint tropism—was significantly diminished in its previously unknown role in PG binding. Our findings provide new insights into B. burgdorferi physiology and pathogenesis, including a previously unappreciated relationship between an immunodominant antigen and the Lyme spirochete sacculus.
Results
Characterizing the function of BB0605 in B. burgdorferi physiology
We sought to understand how the chemical and architectural properties of B. burgdorferi PG may impact Lyme disease pathogenesis. Mur enzymes are essential intracellular proteins dedicated to producing the PG precursor(s) necessary for de novo cell wall (or murein) synthesis and as such, cannot be easily inactivated or altered by genetic manipulation. On the other hand, transpeptidase and transglycosylase enzymes involved in incorporating PG sub-units, and in modifying the existing sacculus, are dispensable under certain environmental conditions [21–23]. We reasoned that this may be the case for B. burgdorferi as well, so we mined a previously published transposon mutant library for integration events that may disrupt transpeptidase and transglycosylase reading frames [24]. Two such clones had transposon integrations at nucleotide positions 152 and 300, relative to the beginning of the bb0605 gene. BB0605 is predicted to encode D-Ala-D-Ala carboxypeptidase since it carries common features shared by PBP5/DacA family orthologues, including 1) a membrane bound, extended amphipathic helix C- terminus; 2) N-terminus that lacks a helix; and 3) a loop shielding the catalytic Serine residue (S1A–S1D Fig, and [25]).
To understand the potential role for bb0605 in B. burgdorferi PG biosynthesis we sought to create an isogenic mutant through typical gene disruption methods [26], but all attempts failed for unknown reasons. Instead, we took an alternative approach whereby the B. burgdorferi 5A18NP1 bb0605::Himar-1 transposon mutant genomic DNA was used as a template for PCR using primers which flanked the target locus by ~2kb. This amplicon was sub-cloned, the resulting plasmid was linearized and then used to displace the bb0605 locus in the B. burgdorferi B31-5A3 genetic background by allelic exchange. The bb0605 locus lies in a putative operon with genes encoding for CheD and a predicted DNA helicase (BB0607, Fig 1A). To assess our new strain, and understand any potential polar effects, we isolated RNA and compared mRNA levels to a parental, B. burgdorferi B31-5A3 control strain [27]. Replacement of the targeted locus with the Himar-1 transposase knockout construct resulted in undetectable amounts of bb0605, while levels in the wild-type strain were comparable to the highly expressed flaB (Fig 1B). Importantly, mRNA levels of the adjacent loci (i.e., cheD and bb0607) were unaffected by our genetic manipulation (Fig 1A and 1B). Transcriptomic analysis of the operon suggested complete disruption of bb0605 expression, without impacting other loci, but we confirmed both the integration site (after nucleotide 299 of the gene) and the absence of distal polar mutations in coding regions of our mutant strain using whole genomic sequencing (WGS) (S1 Table). Together, we contend our new strain (B. burgdorferi B31-5A3/bb0605) is incapable of synthesizing bb0605 and WGS suggest the mutant is free of any unintended mutations.
(A) Schematic of the putative bb0605 operon, as well as the downstream bb0608 locus, mapped from the B. burgdorferi B31 type strain genome. The transposon integration site (+300, 5’ to 3’) in bb0605 is shown (triangle) as are the length, and location, of predicted non-coding regions. (B) Relative mRNA levels of target loci to validate B31-5A3/bb0605 mutant. RT-PCR was performed on RNA isolated from both parental (B31-5A3, charcoal) and mutant (B31-5A3/bb0605, light gray) strains using locus-specific primers and normalized relative to constitutively expressed flaB. Values are the mean + /- SD. n.d., not detected after 45 cycles. (C) Identifying changes in the peptidoglycan composition of B31-5A3/bb0605. Liquid chromatogram of mutanolysin treated PG isolated from B31-5A3 (top, in black) and B31-5A3/bb0605 (bottom, in green). Each numbered peak corresponds to a muropeptide found in Tables 1 and 2, respectively. Peak 5 (shaded gray) corresponds to the muropeptide GlcNAc-MurNAc-Ala-Glu-Orn-[Gly]-Ala-Ala). (D) MS (top) and MS/MS (bottom) of the GlcNAc-MurNAc-Ala-Glu-Orn-[Gly]-Ala-Ala muropeptide. MS/MS fragmentation (inset) data confirms the pentapeptide(-Gly) structure.
Muropeptide analysis of peptidoglycan purified from B. burgdorferi B31-5A3.
The B. burgdorferi genome is riddled with anomalies and functional inferences made purely on silico homology have been misleading in the past [28–30]. As such, we assessed the PG composition of the bb0605 mutant using our recently described high-resolution LCMS method [20]. Our analysis revealed disaccharide-pentapeptide(-Gly)—a structure nearly undetectable in wild-type bacteria—became one of the most predominant muropeptide species in BB0605-deficient bacteria, which we confirmed by reconstructing the fragmentation pattern in MS/MS spectra (Fig 1C and 1D). Additionally, high-molecular weight PG peaks corresponding to cross-linked species were altered in terms of their retention time and abundance (Fig 1C). The latter suggested the formation of new cross-linked species. Further mining of MS/MS data identified many instances in which the new D-Ala-D-Ala species were now present in cross-linked muropeptides, while the same structures did not exist in wild-type PG (Tables 1 and 2). Surprisingly, the degree to which muropeptides were cross-linked in each strain did not change (~33%), but di-D-Ala-containing muropeptides were now present in ~55% of the total amount of peptide bridges (S2 Table). These findings strongly suggest that BB0605 is a functional D-Ala-D-Ala carboxypeptidase. Most importantly, our mutant strain was altered in both PG peptide composition and the type of cross-linked species present in the nascent B. burgdorferi sacculus. Given the homology (S1B–S1D Fig), and the clear biochemical properties of the native enzyme in the B. burgdorferi cell, we will refer to BB0605 as DacA from here forward.
PG synthesis is intimately linked to all aspects of the bacterial cell cycle in most bacteria [31–34]. Dysregulation of any major cell cycle event usually results in gross morphological and/or physiological phenotype(s) [22,35–37]. To assess the overall physiological state of our mutant, we used proxies for cell cycle and cell envelope homeostasis and compared our results to the parental strain. We first assessed the growth by back-diluting cultures and manually enumerated cells daily. Despite exhibiting a longer lag phase, the exponential growth rate of the dacA mutant was slightly faster than the parental control (10.8 vs. 13.6 hours) and both reached a similar final density (S1E Fig). We also assessed both the cell length and the coefficient of variation (CV), which were almost identical: parental CV 35%; mutant CV 34% (Fig 2A). CV of cell length is a proxy for cell cycle homeostasis and can be affected by PG alterations [32,35,38–41]. We further assessed the possible impact of the carboxypeptidase mutant on the B. burgdorferi cell cycle by quantitative analysis of PG synthesis. Analysis of cells pulse-labeled with fluorescent D-Ala, in conjunction with real-time localization studies of known divisome proteins indicate B. burgdorferi elongates by synthesizing discrete zones of PG that are equally spaced and become future sites of cell division [42,43]. We repeated the former and assessed the relative sub-cellular locations of PG synthesis in both strains throughout the cell cycle (i.e., from cell birth to division) and found that the absence of DacA did not affect the position of zonal growth sites (Fig 2B and 2C). During these studies we noticed the fluorescent signal intensity appeared to be higher in the mutant (Fig 2C). We quantified the fluorescent signal from each population by determining the intensity of each pixel in each cell, calculated the average, and normalized by area for hundreds of cells and found the signal in the dacA mutant was ~ 5-fold higher (Fig 2D). Our interpretation of these collective results is that the B. burgdorferi cell cycle is not altered in the absence of DacA and that fluorescent signal intensity differences are likely due to the terminal D-Ala (i.e., where HADA is incorporated) not being processed by the D-Ala-D-Ala carboxypeptidase.
(A) Cell length analysis as a proxy for defects in cell cycle progression. B. burgdorferi B31-5A3 (gray, n = 410) and B31-5A3/dacA (green, n = 724) strains were cultured to mid-log exponential growth, imaged by phase-contrast microscopy, and micrograph data collected was analyzed with Oufti to determine the length of each cell in the population with sub-pixel resolution [88]. Shown are the mean + /- SD, as well as the coefficient of variation (CV). (B) Analysis of zonal PG synthesis. Phase-contrast (upper), Epifluorescence (middle), and merge (lower) images are shown for both strains incubated with 0.25 mM HADA for 4 hrs. Scale bar = 5 µm. (C) Demograph of HADA signal intensity derived from B. burgdorferi B31-5A3 parental (top, n = 410) and B31-5A3/dacA mutant (bottom, n = 724) strains, attained from individual cells, organized by length. Signal intensity is shown in arbitrary units (A.U) as a heatmap. (D) Population-level HADA fluorescent signal intensity analysis of data collected in C, B. burgdorferi B31-5A3 (gray) and B31-5A3/dacA (green). **** p < 0.0001, unpaired Student’s t-test. (E) Swarm assay to assess motility. A single soft agar plate was inoculated with 7.5 µL of 109 cells/mL suspensions of both strains B. burgdorferi B31-5A3 (gray) and B31-5A3/dacA (green), on opposite sides, and colony diameters were measured after 5 days. Data shown are the mean of 3 replica plates, + /- SD. Statistical significance was assessed using an unpaired Student’s t-test.
Motility is essential to B. burgdorferi infectivity and pathogenesis [44]. Given the conceptual and biophysical relationship between the periplasmic flagella and B. burgdorferi PG [20,45,46], we considered the possibility that our mutant strain with altered PG organization may have motility defects. We did not detect any significant difference between strains using swarm plate assays, suggesting the PG peptide composition and linkage status did not alter the intimate relationship between the two distinct but related B. burgdorferi ultra-structures (Fig 2E). Together, our mutant strain characterization indicates that the B. burgdorferi DacA homologue is a bona fide terminal PG peptide carboxypeptidase, and that alterations created as a result of its inactivation do not produce gross cellular or physiological defects (Figs 1, 2, and S1E).
Altering PG peptides attenuates Lyme arthritis
B. burgdorferi PG has been implicated in the development of Lyme arthritis in both mice and humans [13,14]. To assess how changes to PG structure may impact the former, we needle inoculated a cohort of C3H mice with either the dacA mutant or parental control strain (105 cells) and tracked arthritis severity for > 12 weeks. C3H were selected because of their inherent capacity to develop severe murine Lyme arthritis [47–49]. Wild-type B. burgdorferi B31-5A3 produced visible arthritis in C3H mice within 2 weeks post infection, while the dacA mutant appeared to lag (Fig 3A and 3B). Five mice were sacrificed at this time point and while subtle differences were apparent in bacterial load, as assessed by qPCR, none were statistically significant (S2 Fig). Wild-type infected mice continued to experience marked arthritis, which peaked 28 days after inoculation and remained apparent for at least 60 days post-infection by visual metrics (Fig 3A–3C). Blind histopathological analysis—a more quantitative assessment of arthritis—was consistent with gross visual arthritis scoring throughout wild-type infection (Fig 3B and 3D). Infection with mutant bacteria unable to produce DacA, on the other hand, produced undetectable levels of murine Lyme arthritis in virtually all mice (Fig 3). A single animal (4%) appeared to have histological evidence of arthritis, but severity did not increase in the population of mice infected with B. burgdorferi B31-5A3/dacA and remained near basal levels even after 12 weeks (Figs 3A, 3B, 3D, and S3).
Mice injected with B31-5A3/dacA exhibit severely attenuated Lyme arthritis. (A) Both wild-type B31-5A3 and B31-5A3/dacA bacteria were needle inoculated (105 cells) into C3H/HeJ mice (n = 25 per group), and arthritis severity was assessed by visual arthritis scoring every two days. Two and four-weeks post injection, five and ten animals, respectively, were sacrificed (dotted lines) for histopathological analysis and qPCR (see Figs 4 and S4). Mean + /- SEM are shown. (B) Representative histopathology of HE-stained sections of tarsal joints from mice infected with B. burgdorferi B31-5A3 (left) and B31-5A3/dacA (right) 14-, 28-, and 84-days post injection. A single PBS control tissue is also shown. Arthritis scores for tissues shown in this panel were as follows: * = 0.0, ** = 0.25, # = 1.0, ## = 2 – 3. (C) Representative ankles from C3H/HeJ mice infected with B. burgdorferi B31-5A3 (upper) or B31-5A3/dacA (lower), 4-weeks after infection. (D) Blind histopathological scores for individual mice in each group at 14-, 28-, and 84--days post PG administration. Statistical significance was determined using Welch’s t-test, * p ≤ 0.05 (*).
Disseminated stages of Lyme disease in mice produces a multi-system infection whereby bacteria can be cultured from, and their products detected in, various tissues [50,51]. We first determined that the significant attenuation of murine Lyme arthritis was not simply due to the dacA mutant strain being non-infectious, which was confirmed by culturing live bacteria from the bladders of all mice, infected with both strains (Table 3). We returned to qPCR of select tissues to further understand the arthritis phenotype and determine if bacterial load was playing a role. We focused on three tissues that frequently harbor spirochetes in murine Lyme disease—heart, ankle, and skin. The latter was assessed in ear tissue, which is distal relative to the sub-cutaneous site of mouse inoculation (rear flank) and thus an excellent measure of bacterial dissemination. Four-weeks post infection, both skin and heart tissues from mice infected with both strains contained comparable levels of B. burgdorferi DNA (Fig 4A and 4B). Quantitative assessment of ankle tissues, however, resulted in a modest, but significantly lower bacterial burden in animals infected with the dacA mutant (Fig 4C). Total IgG levels at this stage of infection were similar (Fig 4D). We also assessed ankle tissues 12 weeks post infection, a point at which arthritis was still visible in wild-type infected animals and found no difference between strains (S4 Fig). These data suggest PG composition and overall architecture are important determinants of Lyme arthritis, findings that may be partially explained by playing direct, or indirect, role(s) in tissue tropism.
(A) qPCR analysis of heart (A), ear (B), and ankle (C) tissue 4-weeks post injection. The number of copies of recA were quantified using a standard curve and then normalized to murine nidogen. The black bars indicate the average normalized bacterial load for each group of tissues, + /- SD. Statistical significance was calculated using an unpaired Student’s t-test p ≤ 0.01 (**). (D) IgG response to infection. Total, B. burgdorferi-specific IgG levels were quantified by incubating plasma samples from mice injected with B31-5A3, B31-5A3/dacA or PBS in 96-well microplates coated with B. burgdorferi lysate and determining absorbance at 450 nm.
p83/100 antigen is a peptidoglycan-associated protein
PG is often directly or indirectly associated with proteins which 1) provide cell wall stability; 2) secure the position of the sacculus relative to other cell envelope components; 3) ensure continuity between envelope layers; and 4) are occasionally involved in bacterial pathogenesis [52–54]. We previously discovered that Neutrophil Attracting Protein A (NapA) was directly associated with the B. burgdorferi PG sacculus and seemingly served many of the functions described above. During our assessment of PG-associated proteins (PAPs) in the Lyme spirochete, apart from NapA, the next most likely PAP candidate was the immunodominant protein p83/100 (BB0744) [15,55–58]. Despite being highly expressed and a serological indicator of infection, little is known about the biological function of p83/100. Intriguingly, mutant bacteria unable to produce p83/100 exhibited defects in tissue tropism and tracking studies in live animals found a paucity of bacteria in infected animal joints [59]. The same study also found that p83/100 localized to the periplasm of B. burgdorferi. In the context of the current findings, these data suggested that p83/100 might be a PAP, and that altered PG peptides (e.g., those present in the dacA mutant), may impact this association. We therefore assessed this possibility by purifying PG from both wild-type and dacA mutant bacteria by boiling each in 5% SDS. After removing SDS we treated half of the PG sacculi preparation with proteinase K (PK) and left the remainder untreated. We also performed the same procedure with a p83/100 mutant strain, as an additional control, and developed an ELISA approach to detect p83/100. A titration of each purified PG preparation was immobilized on microtiter plates using an anti-PG antibody (r-mAb-2G10, [14]), and probed using anti-p83/100 or biotin conjugated r-mAb-2G10. The latter was used to normalize for the total amount of sacculi present in each preparation since slight variations are common due to the lengthy purification process. Purified sacculi isolated from wild-type bacteria contained p83/100-derived fluorescent signal, which increased as the amount of untreated (i.e., PK-) cell wall-containing material increased (Fig 5A). Following protease treatment (PK+), this same sample contained near undetectable levels of p83/100, regardless of the titration (Fig 5A). Purified PG from dacA mutant bacteria contained significantly less p83/100 and the p83/100-derived signal returned to basal (i.e., p83/100 mutant) levels after degradation by PK (Fig 5A). Immunoblots with lysates from each strain suggested that cellular levels of p83/100 were comparable and similar to the highly expressed outer surface protein OspA and the FlaB loading control, regardless of DacA production (Fig 5B), indicating that the lower levels of PG-bound p83/100 in the dacA mutant is not due to overall expression.
(A) ELISA-based approach to assess PG-p83/100 interaction(s). Sacculi from B. burgdorferi B31-5A3, B31-5A3/dacA, B31-ML23/p83/100 strains were purified and one-half was treated with proteinase K (+PK) while the other was incubated with buffer (-PK). A serial titration of each purified PG preparation was immobilized on microtiter plates using r-mAb2G10, and the amount of PG and p83/100 was quantified using biotin-r-mAb2G10/streptavidin:HRP, anti-p83/100 coupled with anti-rat:IgG:HRP, respectively. The PG value attained for each strain and preparation was used to normalize each p83/100 value, which were plotted as a function of absorbance (450 nm). (B) Lysates from each strain were prepared and total protein loading was optimized using immunoblots with anti-FlaB. After optimization, each immunoblot was performed with the exact same amount of lysate to detect putative PAP (anti-NapA, anti-p83/100), as well as Outer surface protein A (anti-OspA), and the FlaB loading control was repeated (anti-FlaB). (C) The same samples described in ‘A’ were used to assess the amount and localization of p83/100 in purified PG. Each PG preparation and treatment were immobilized on optical glass using poly-Lys and probed using PG-binding lectin Wheat Germ Agglutinin conjugated to Oregon Green 488 (PG, green, upper panel) and anti-p83/100 conjugated to anti-rat IgG:Alexa647 (p83/100, magenta, middle panel). A merge of both micrographs described above is also shown using the color scheme described above (lower panel). Scale bar 5 µm. (D) Total integrated p83/100 signal intensity for each sample and preparation described in ‘A’ and ‘C’. The signal intensity of each pixel, inside each sacculus was measured, and the average was calculated for each, then normalized by area (one dot). This was repeated for > 100 sacculi per sample and the median of each is shown (black line). One-way analysis of variance was used to determine significance between P83-derived signal present in B31-5A3 wild-type sacculi (-PK), relative to B31-5A3/dacA sacculi (-PK), **** p < 0.0001, ANOVA. Note that all other comparisons to wild-type (-PK) sacculi were also statistically lower (****) but not shown for simplicity. n values for each were as follows: B31-5A3 (-PK), 227; B31-5A3 (+PK), 132; B31-5A3/dacA (-PK), 255; B31-5A3/dacA (+PK), 133; B31-ML23/p83/100 (-PK), 178; B31-ML23/p83/100 (+PK), 101. (E) Population-level line scans of values attained in studies described in ‘C’. P83-derived signal intensity present in purified sacculi from B31-5A3 (-PK, gray) and B31-5A3/dacA (-PK, green) was determined by measuring the signal intensity of each pixel along the midline of the cell wall, for each sacculus, and plotted as a function of normalized length (0 – 1). Black lines show the mean of each, SD are shaded. (F) Demograph analysis of p83/100 signal intensity in sacculi purified from B31-5A3 (-PK).
Our initial molecular assessment provided clear evidence that p83/100 is a PAP and alterations to the chemical properties of PG significantly diminish putative cell wall-protein interaction(s) (Fig 5A). We next assessed the same PG samples described above using immunofluorescence to provide additional cellular and spatial context. Sacculi were immobilized on optical glass and probed using anti-p83/100:anti-rat-Alexa647. To prevent any ambiguity in our initial results, we visualized sacculi using fluorescently conjugated Wheat Germ Agglutinin rather than r-mAb2G10. Prior to PK treatment, sacculi isolated from wild-type cells show a clear p83/100 signal, which was barely visible in the dacA mutant PG (Fig 5C). To further assess the reduction in signal we measured the intensity of each pixel, inside each sacculus, calculated an average, and normalized the values by total area of the object. We then repeated this procedure, in an automated fashion, for more than 100 sacculi per sample and treatment. Our population-level results suggest that, on average, dacA mutant sacculi contain six times less p83/100 (Fig 5D), confirming our initial ELISA results. p83/100 signal distribution was assessed at the population level using quantitative line scans and demographs; both yielded similar results and suggested some periodicity throughout the long axis of the sacculus as well as enrichment at the PG poles (Fig 5E and 5F). In summary, our data indicate that p83/100 is a PAP whose interactions are impacted by PG-peptide composition and both p83/100 and PG together affect bacterial tissue tropism and the development of Lyme arthritis.
Discussion
Lyme arthritis is an immune-mediated, pathophysiological consequence of B. burgdorferi infection. B. burgdorferi PG has been implicated as a mediator of the maladaptive response by potentiating sustained inflammation [8,13,14]. One of the many features that differentiate the Lyme disease spirochete cell wall from others are its unusual components, which are likely drivers in arthritis [14]. The initial impetus of this work was to manipulate the overall PG architecture to directly address this scenario in the context of a natural infection, which led to studies on a dacA mutant strain. While we were completing this study, McCausland et al. published their findings using the original transposon mutant of dacA. Despite differences in the motivation to assess the mutant, as well as the genetic background and the approaches used, they came to the same conclusions in terms of the putative function of BB0605 as a structural homologue of DacA [60]. The same study found dacA mutant bacteria have marked increase in penta-peptide-(Gly), produce a central zone of PG synthesis similar to wild-type cells, and exhibit higher signal intensity when incubated with a fluorescent analogue of D-Ala [60]. Robust agreement in the biochemical and cellular data between groups further supports the overall function of BB0605 as a carboxypeptidase and bolsters our additional findings, which we discuss in the context of B. burgdorferi physiology and pathogenicity below.
The bacterial kingdom is full of genotypic and phenotypic diversity; production and activity of PBPs contribute to both. Most bacteria produce 10–16 discrete PBPs. In silico and activity-based profiling studies suggest B. burgdorferi encodes 4–5 PBPs and DacA is the only predicted type-5 PBP [10,42,43]. PBP5 homologues are known as low molecular weight (LMW) enzymes that typically lack transpeptidase activity, but are often important in controlling cell shape, as well as in ensuring faithful completion of the bacterial cell cycle [25]. For instance, both morphological and cell envelope abnormalities are common in E. coli mutants unable to produce DacA [61–64]. Like B. burgdorferi, many gram-positive bacteria produce a single PBP5 orthologue, and they too are directly linked to both cell shape and division [65,66]. Morphological variability is attributed to uncoupling PG precursor processing from nascent synthesis leading to osmotic vulnerabilities. Aberrant cell cycle events in carboxypeptidase mutants are thought to be the result of the divisome machinery acting on a surplus of di-D-Alanine-containing PG peptides at the septum [25]. As such, it was surprising to us that dacA deficient B. burgdorferi 1) did not exhibit a growth defect; 2) appeared morphologically identical to wild-type bacteria; 3) produced properly placed zones of PG synthesis; and 4) formed a viable septum at mid-cell (S1 and S2 Figs). It is important to note that many of the phenotypes associated with PBP5 mutants are often context-dependent and require additional mutations in other PBPs and/or exogenous stress; conditions that may be more reflective of in vivo environments [67].
PG does not appear to play an active role in B. burgdorferi shape, but it does counteract deforming torsional stress imparted by endo-flagella wrapping around the cell cylinder [44,68]. Thus, it is thought that the passive function of PG in motility is to provide reciprocal elastic forces that are fine-tuned to create the least amount of drag and allow the flagella to produce the greatest amount of torque [20,46]. Unlike other helical bacteria, whose shape and motility are impacted by carboxypeptidases and cross-linking [69–71], B. burgdorferi motility in vitro and dissemination in vivo are unaffected in a dacA mutant (Figs 2E, 4A, and 4B). We postulate B. burgdorferi has evolved alternative strategies to maintain PG flexibility and continuity that go beyond cross-linking, likely through glycan modifications [20]. It is also possible that maintaining the terminal D-Alanine in cross-bridges do not confer a substantiative biophysical effect but may alter the cell envelope in other ways.
There are numerous examples by which carboxypeptidases are directly responsible for beta-lactam resistance; often through allelic variation and altered cellular concentrations of PBP5 homologues [72–75]. Assessing the true pathogenicity of unprocessed PG stem peptide—resulting in increased pentapeptide and altered cross-links—is challenging because it is difficult to uncouple the fitness cost(s) from the alteration. Yet, some instances are worth noting. Carboxypeptidase Pgp2 is necessary to establish Campylobacter jejuni infection in a chick colonization model [71]. DacA-deficient Streptococcus pneumoniae remains infectious, but target cell adherence is impaired, and mutant bacteria are more likely to die following phagocytosis [76]. The DacA orthologue in Staphylococcus aureus, LMW PBP4, is essential for invasion and colonization of osteocytes, a pre-requisite to causing osteomyelitis, likely by altering the proinflammatory capacity of S. aureus PG [77]. It appears B. burgdorferi DacA may be functionally similar to S. aureus PBP4 in the context of pathogenesis, since similar sacculi changes result in the near complete attenuation of Lyme arthritis, a pathological manifestation of Lyme disease to which PG has been directly linked (Figs 1 and 3, and [13,14]).
We found p83/100—a periplasmic protein of unknown function, also known as BB0744—is a PAP and the carboxypeptidase activity of B. burgdorferi DacA directly or indirectly impacted the PG-protein interaction. Evidence for p83/100 interactions with B. burgdorferi PG were initially provided in an unbiased proteomics approach [15]. We serendipitously returned to these published data when we observed apparent tissue tropism in the dacA mutant strain (Fig 4 and see below) and critically assessed p83/100-PG interactions through molecular and cellular approaches (Fig 5). Further support for this association is provided by earlier studies in which p83/100, like PG and PAP NapA, is also present in B. burgdorferi outer-membrane vesicles [15,78]. PG-p83/100 complexes are near ubiquitous throughout the sacculi, with some enrichment at the PG poles, while p83/100-derived signal in intact cells appears uniform (Fig 5E and 5F, and [59]). It is possible that the increased polar signal in purified sacculi reflects the static, PG-linked population of p83/100 in the cell, which is not captured when visualizing the entire periplasmic space. It remains to be determined how p83/100 interacts with the B. burgdorferi cell wall, but resistance to boiling SDS treatment suggests a covalent linkage. It is possible that the interaction(s) are non-covalent and could be released with longer SDS treatment, as was the case for LipL21 in Leptospira interrogans [79]. We favor a model whereby interactions involve cross-linked PG peptides and the presence of an additional D-Ala results in steric hinderance, which impact the PG-p83/100 interaction(s). This scenario is somewhat reminiscent of Braun’s lipoprotein (Lpp), which is covalently attached to the alpha carboxylic acid through the L-center of mDAP [80,81]. The primary function of Lpp is to be a positional scaffold for the E. coli PG sacculus, relative to the outer membrane. The distance between the inner and outer membranes of B. burgdorferi is ~ 20 – 40 nm [13,37]. If the extended N-terminal alpha-helix in the predicted structure of p83/100 (~ 8 nm) along with a C-terminal membrane-associated amphipathic alpha-helix are correct, then one could envision it functioning as a tether between the sacculus and the outer membrane (S5A Fig). This could occur through protein-protein interactions in the outer membrane, or in the periplasm, but it is this unknown protein that is surface exposed and primarily responsible for mediating tissue residency in interarticular spaces and/or periarticular vascular beds (S5B Fig and see below).
Lyme disease-causing Borreliae spp. are known to exhibit tissue tropism, leading to diverse and distinct pathologies [5]. Initial studies of an isogenic mutant of p83/100 indicated that ear and heart tissues were devoid of viable bacteria 3 weeks post-infection, as well as a marked reduction in bacterial load in infected ankle joints [59]. We found that the dacA mutant, which possesses less PG-bound p83/100 (Fig 5A–5C), had no apparent difference in ear or heart burden, but bacterial DNA in mouse ankles was significantly reduced (Fig 4). A simple explanation for this subtle discrepancy is the time point in which each were assessed (3 weeks versus 4 weeks post-infection, Fig 4). We also noticed differences in bacterial load 2 weeks post-infection, but none were significant (S2 Fig). Together, we speculate that the PG-p83/100 complex is important in stabilizing the cell envelope and that this association becomes critical in the viscous joint space environment. We also argue that p83/100 alone cannot mediate Lyme arthritis since joint inflammation is still attenuated when bacterial burden is similar (Figs 3, 4, and S4). Thus, we favor a model in which both PG-p83/100 interactions (along with unknown, surface-exposed protein(s) (i.e., tissue and disease tropism)) and PG architecture (inflammatory mediator of arthritis) are key players in the murine Lyme arthritis arena. In other words, p83/100 functions as a PAP that may indirectly impact the fitness of the spirochete in the joint environment by affecting the tissue-dependent context of outer surface protein expression while the chemical constituents and organization of the PG sacculus directly contribute to its inflammatory capacity [82,83]. Our attempts to uncouple these biologically linked processes (i.e., phenotypic change in PG chemistry, P83/100 interaction, and presence of DacA) by complementing the dacA mutant with either a wild-type or catalytically inactive allele have failed for unknown reasons, but these studies will be important to improve our understanding of this system and identify causal relationships. It will also be intriguing to determine if the synovial fluid of human Lyme arthritis patients, which contains PG and potentially other PAPs, has detectable levels of PG-p83/100 complexes as well [13–15,84].
The bespoke connection between the unusual B. burgdorferi PG, and the equally irregular Lyme spirochete cell envelope, continue to be elucidated. Conventional reasoning posited that B. burgdorferi does not produce PAPs. We now have three PAP candidates that are intertwined in the physiology, life cycle, and pathogenesis of B. burgdorferi (NapA, p83/100, and BpiP [85]). Further assessment of these proteins, their interactions, and the coordination thereof, will certainly further our understanding of Lyme disease and may provide new strategies to prevent or mitigate human illness.
Materials and methods
Ethics statement
All animal studies were approved by the Virginia Tech Institutional Animal Care and Use Committee (Protocol numbers 20-051 and 23-099).
Bacterial strains and culture conditions
All strains used in this study were derived from the B. burgdorferi type strain B31 [10] The primary strain for all experiments was the infectious B. burgdorferi B31-5A3 [27], kindly provided by Dr. Jenifer Coburn (Medical College of Wisconsin), from which the B31-5A3/dacA strain was derived as described below. The p83/100 (locus tag bb0744) mutant strain—B. burgdorferi B31-ML23/bb0744—was graciously provided by Dr. Jon Skare [59]. All strains used in this study were cultured in Barbour-Stoenner-Kelly II (BSK-II) medium, pH 7.6, supplemented with 6% heat inactivated rabbit serum (Pel-Freez Biologicals) at 34°C unless stated otherwise [86]. Organ outgrowth studies were conducted using the same media described above but supplemented with a final concentration of 2% of gelatin, and amphotericin (2.5 μg/mL), fosfomycin (20 μg/mL), and rifampicin (50 μg/mL). When necessary, the culture media was further supplemented with gentamicin (B31-5A3/dacA, 40 µg/mL) or streptomycin (B31-ML23/bb0744, 100 µg/mL).
To assess any potential growth defects, BSK-II (5 mL) was inoculated with 5 x 105 cells and bacterial density was determined in triplicate every 24 hours using a 0.02mm Petroff-Hausser hemocytometer (Hausser Scientific) [87]. Data were fit to a log-linear model to determine the exponential growth phase and these values were used to calculate growth rate.
Swarm plate assay
To assess whether the change in PG structural motifs has any impact on the motility of the B31-5A3/dacA strain, we prepared semi-solid medium plates as described previously with a final concentration of 0.5% (w:v) of low-melt agarose [20]. B31-5A3 and B31-5A3/dacA cultures (15 mL, 34°C) were harvested at mid-log and resuspended to a concentration of 109 cells/mL. Three plates were inoculated on both sides with 7.5 µL of either suspension. After five days of incubation at 37 °C and 5% CO2, the radius of each colony was measured in five different directions and averaged.
Genetic manipulation of B. burgdorferi and whole genome sequencing
Clone T08TC493 was previously described and contains a Himar1 transposon insertion event 300 bp downstream of the predicted bb0605 start site [24]. DNA primers were created (F: 5’-GGTCCAAAAAGAACATTTGAAACCG-3’; R: 5’-TCTTCTATGCTTTCAATGACCTGCT- 3’) to amplify Himar1 and ~2kb of flanking DNA and the product was sub-cloned (TOPO TA pCR2.1). The resulting plasmid was purified, sequenced, then linearized and used to transform B. burgdorferi B31-5A3 by electroporation, and recombinant clones were selected as described (Samuels 1994). For whole genome sequencing, bacteria were cultured in 40 mL of BSK-II, harvested by centrifugation (4,000 x g for 10 mins) and washed with PBS before genomic DNA purification using Zymo Research Quick-DNA Miniprep kit following manufacturer’s recommended procedures. Whole genome sequencing was performed by the Microbial Sequencing and Analysis Center at Pittsburgh University (Pittsburg, PA), which resulted in > 410X coverage for each sample. Unicycler processed and assembled all sequence data, which was mined for polymorphisms and the Himar1 integration site. Data were also mined to determine the plasmid profile of each strain, which was the same, and identical to previously published for our B31-5A3 clone [20]. One sequence confirmed clone was named B. burgdorferi B31-5A3/bb0605 (or/dacA) and the focus of our studies.
PG purification and analysis
0.25–0.5 L of mid-log cultures were subject to isolation procedures established previously [13,20,42]. Briefly, following mutanolysin-mediated peptidoglycan digestion, the isolated supernatant containing soluble muropeptide fragments was frozen and dried using a high vacuum line coupled with a liquid nitrogen solvent trap. The purified muropeptides were resuspended in 150 µL of saturated sodium borate buffer, pH 9.25. A 50 µL aliquot of 100 mg/mL NaBH4 was added dropwise and after 1 hr incubation the reaction was quenched with approximately 10 µL of formic acid (final pH 3). The samples were then snap frozen and dried once more with a high vacuum line. The dried muropeptides were resuspended in H2O:MeCN (200 µL, 9:1, v:v) containing 0.1% formic acid and sonicated in a water bath for 10 min followed by centrifugation at 13,000 x g at 4°C for 10 min. 180 µL of the supernatant was transferred to a LCMS vial for analysis.
Microscopy
To determine whether zonal growth was impaired, both B31-5A3 parental, and B31-5A3/dacA) mutant strains were cultured to mid-log (107 cells/mL) and incubated with 0.25 mM HADA for 4 hours, as previously described [42]. The cultures were then centrifuged at 3,500 x g for 15 minutes and washed three times with PBS. The final pellets were then resuspended in 100 µL of PBS and 5 µL of these suspensions were spotted on PBS agarose (2%) pads prior to imaging at RT with a Zeiss Axio Observer equipped with an oil-immersion phase-contrast objective Plan Apochromat 100 × /1.45 numerical aperture (Nikon) coupled to a Hamamatsu Orca-Flash 4.0 V3 Digital CMOS camera. Automated cell detection and cell-length measurements were facilitated by Oufti [88]. Demograph, signal intensity scatter plots, and population-level line scan analyses involved custom MatLab scripts, published elsewhere [15,42,43].
Mouse studies
Three- to four-week-old C3H/HeJ mice were purchased from Jackson Laboratory. To manage the experimental procedures and data collection associated with the number of mice in this study, all data collected and shown are the results collated from two independent experiments. Both strains of B. burgdorferi were cultured to mid-log (5x107 cells/mL) and resuspended to a concentration of 1x106 cells/mL in PBS following centrifugation at 3,500 x g and 4 washes with PBS. Mice (n = 25 per strain of B. burgdorferi) were subcutaneously inoculated with either strain between the scapulae with 100 µL of this suspension; an inoculum of 1x105 cells.
The mice were euthanized through CO2 asphyxiation followed by cervical dislocation and heart puncture at 2, 4, and 12 weeks post-inoculation. Tissues including the heart, ears, bladders, and tibiotarsal joint were harvested for outgrowth culture, qPCR quantification of bacterial load, and histopathology. Outgrowth cultures were monitored by dark-field microscopy for the presence of viable cells. Cultures were considered positive if motile spirochetes were present 4–6 weeks post inoculation.
Arthritis scoring
All mice were examined on alternate days for signs of hindleg swelling and inflammation and assigned a score based on the severity of visible signs of arthritis using a scale and method previously described [47,48]. Briefly, no swelling of the carpal and tarsal joint or redness of the paw is a score of 0; some swelling accompanied with redness of the paw is a 1; moderate swelling, redness of the paw is a 2; severe swelling, redness of the paw, and occasional swelling of the interphalangeal joint is a 3.
Histopathology and semiquantitative analysis
Rear limbs with the skin removed and previously immersion-fixed in 10% neutral buffered formalin were received blind to experimental manipulation by the Comparative Pathology Research Core (CPR Core), Department of Comparative Medicine, Yale University, School of Medicine, New Haven CT. The limbs were decalcified using Decal Solution (Richard Allen Scientific, Kalamazoo, MI) then trimmed and placed medial side down into tissue cassettes by routine methods. The tissue cassettes were submitted to the CPR Core Histology Laboratory for processing, embedding, sectioning and staining by hematoxylin and eosin (HE) and cover-slipped by routine methods.
The tissues were examined blind to experimental manipulation using the standard Barthold Lyme tenosynovitis (arthritis) semiquantitative scoring system [89]. Briefly, the knee and tibiotarsal joints were scored for arthritis severity on a scale of 0 (negative) to 3 (severe) in a blinded fashion as described previously by CJB [90–92]. The slides were examined using an Olympus BX53 light microscope, photographed using an Olympus DP 28 camera and Olympus cellSens Standard 4.2.1 software. The images were optimized and figures made using Adobe Photoshop 26.8.1.
qPCR
DNA was isolated from murine tissue samples using the Quick-DNA Miniprep Plus Kit (Zymo Research). B. burgdorferi recA and mouse nidogen standard curves were generated using DNA isolated from B31-5A3 grown in culture and mouse tissue, respectively. These were quantified spectrophotometrically, from which the number of genome copies were determined [93].
qPCR was performed with Apex GREEN qPCR Master Mix (Genesee Scientific) on a CFX Opus 96 (BioRad) with the following parameters: 95°C for 15 minutes followed by 40 cycles (95°C for 15s, 59°C for 30s, 72°C for 30s). The primers used to amplify B. burgdorferi-specific recA were F: 5′-GTGGAT CTA TTG TAT TAG ATG AGG CTC TCG-3′ and R: 5′-GCC AAA GTT CTG CAA CAT TAA CAC CTA AAG-3′. The primers used to quantify mouse nidogen were F: 5′-CCA GCC ACA GAA TAC CAT CC-3′ and R: 5′-GGA CAT ACT CTG CTG CCA TC-3′. Data was analyzed through comparison to the standard curves and normalization to the reference gene [93,94].
RT-PCR
Mid-log phase exponential, 40 mL cultures (107 cells/mL) of B. burgdorferi B31-5A3 and B31-5A3/dacA were harvested at 3,250 x g at 4°C for 15 minutes and washed twice with PBS. RNA was purified and processed immediately using Zymo Quick-RNA Miniprep Plus Kit (Zymo Research) using the manufacturers recommended procedures for bacterial cells with one modification. After column purification, each RNA sample was treated with DNaseI for 20 minutes, re-purified, and then digested again with DNaseI (i.e., two 20-min digestions). After final purification, ‘qPCR’ was performed on each DNase-treated RNA sample to ensure preparations were free of DNA using recA primers listed above.
RT-PCR studies to assess expression of each target locus were performed on purified RNA fromB. burgdorferi B31-5A3 and B31-5A3/dacA with the following primer pairs: bb0605 (F: 5’- CTG GAT ATA GCA GCG AGA ATA AG-3’, R: 5’-GCA GTT CCT AAA TTT CTA CTC TTT GG-3’); cheD (F: 5’-CTT GGT TCT TGT GTT GCT GTT GTG C, R: 5’- CTT CCC CTT TGA TCA GGA GAT TAG TC-3’); bb0607 (F: 5’-GCT TTA TGA AGA GAG GTT AAG GC, R: 5’-CTG ATC ATC ATC CCC AAC ACA ACA C-3’); bb0608 (F: 5’- CCA ATT GAA ATT GTT GAA GAG GAT G-3’, R: 5’-CCC CAA GAT GAT AGC TAC TCC AAT C-3’). All reactions were performed on the same plate, using PCRBIO 1-Step Go RT-PCR Kit (PCR Biosystems) following the recommended procedures. Delta Cq values were acquired for each sample and target, then normalized by values attained by the constitutively expressed flaB. Melting curve analysis of each amplicon confirmed a single species was produced for each primer pair. The normalized, relative values attained were plotted from three technical replicates. We note bb0608 was initially included into our analysis because we thought the locus was part of the same reading frame. After further examination, there appears to be sufficient space for an independent promoter, and our data suggest this promoter is not active in vitro (Fig 1B).
Serology
B31-5A3 lysate was diluted to 2 µg/mL (Bradford assay) in carbonate coating buffer (90 mM NaHCO3 and 60 mM Na2CO3, pH 9.6), coated onto 96-well plates and incubated overnight at 4°C. After blocking with 100 µL SeaBlock (Thermo-Fisher) at 37°C for 2 hours, serially diluted mouse blood plasma samples (100 µL/well) were added to the wells and incubated at RT for 1 hour. Afterwards, HRP-conjugated Goat anti-Mouse IgG (1:25,000) was added to each well and incubated for an additional hour. Finally, 100 µL of 1-Step TMB ELISA Substrate Solution (Thermo Scientific) was added and incubated for 10 minutes. The reaction was stopped with 100 µL 1.5 N sulfuric acid and the end-point absorbance at 450 nm was measured for avidity determination.
LCMS analysis
Analyses were performed as published previously [20] on a Shimadzu LCMS9030 QToF system coupled to a LC-40B X3 UPLC, a SIL-40C X3 autosampler (10°C), and a CTO-40C column oven (40°C). Borohydride-reduced mutanolysin-digested muropeptides were separated on a Waters BEH C18 column (2.1 x 50 mm, 1.7 µm particle size), at a flow rate of 0.4 mL/min. Solvent A was water and Solvent B was MeOH, with both solvents containing 0.1% formic acid (v/v). The initial solvent condition was 99:1 (A:B) for 3 min. This was followed by stepwise linear gradients first to 8% Solvent B (12 min), then to 20% Solvent B (24 min), and finally 95% Solvent B (25 min), followed by a hold at 5% B for an additional 5 mins (30 min total). The column was then re-equilibrated to 1% B (1 min of 1% B, followed by a 5 min re-equilibration time). The reduced muropeptides were subjected to electrospray ionization in positive ion mode using data dependent acquisition. Interface voltage was 4.0 kV at 300°C with a desolvation temperature of 526°C and a DL temperature of 250°C. Gas flows for nebulizing, heating, and drying gases were 2, 10, and 10 L/min, respectively. The m/z range for the precursor MS scan was 400–2000, with a fragmentation data range of 50–2000 and a collision energy of 35 ± 10 V.
Western blots
Briefly, 50–100 mL of each strain was cultured to mid-log phase, pelleted by centrifugation at 3,500 x g for 15 minutes, and washed four times with PBS. Washed cell pellets were resuspended in 0.5 mL of PBS, sonicated, and the resulting crude lysate was stored at -80°C. Equal amounts of sonicated lysates, validated by a Bradford assay (Thermo Scientific), were separated by protein gel electrophoresis on SurePAGE Bis-Tris 4–20% precast gels (GenScript) and transferred onto PVDF membranes (Genesee Scientific). Immunoblots involved probing membranes with rabbit or rat antisera against FlaB (1:1,000), NapA (1:5,000), and p83/100 (1:5,000), which were kindly gifted by Melissa Caimano (UConn), Frank Gherardini (NIH), and Jon Skare (Texas A&M University), respectively. Western blots to detect OspA used antibodies (1:1,000) purchased from Rockland Inc. Rabbit IgG:HRP (Jackson Labs) was used to detect anti-NapA, and anti-OspA at dilutions of 1:12,000, and 1:16,000, respectively; rat IgG:HRP (Invitrogen) was used against anti-FlaB and anti-p83/100 at a dilution of 1:8,000. The membranes were incubated with SuperSignal West Pico PLUS chemiluminescent substrate (Thermo Scientific) and imaged with an Odyssey M (LI-COR).
PAP studies
PG was purified from 250 mL bacterial cultures, as described above, but stopped prior to chymotrypsin step. Prior to protease digestion, sacculi were resuspended in 500 µL of PBS and split in half. One 250 µL sample (-PK) was stored at 4°C while the other 250 µL sample was first treated with alpha chymotrypsin, as described previously [15,20], then 10 µg/mL of proteinase K overnight. Proteinases were inactivated and PG purified afterwards, as described above.
Proteinase treated and untreated sacculi, from each strain, were assessed using ELISA and Immunofluorescence. ELISA studies involved two separate procedures: PG quantification and p83/100 quantification. The former was performed as reported elsewhere [14]. p83/100 ELISA involved coating microtiter plates with 1 µg/mL of r-mAb2G10 to immobilize each PG preparation. Unbound material was removed with three washes of PBS-T (PBS + 0.05% Tween-20), and sacculi were blocked for 2 hours at 37°C using SeaBlock (Thermo Fisher). Excess blocking agent was removed with 2 washes of PBS, and sacculi were probed for p83/100 using rat anti-p83/100 serum (1:25,000). Non-specifically bound antibody was removed with 4 washes of PBS-T, anti-rat IgG:HRP (Jackson labs, 1:30,000), was used to detect the p83/100 polyclonal antibody, which was followed by 4 additional PBS-T washes and TMB one-step ultra (Thermo Fisher) substrate for colorimetric enzyme detection. Immunofluorescent studies were performed exactly as described previously [15], but rat anti-p83/100 (1:1,000) and rat anti-IgG:Alexa647 (1:400) antibody pairs were used.
Statistical analysis and modeling
All statistical analyses are described throughout the text and figure legend, which were performed using GraphPad Prism 10. AlphaFold 3.0 was used for all protein structural modeling [95].
Supporting information
S1 Table. Summary of whole genome sequence data of dacA mutant strain.
https://doi.org/10.1371/journal.ppat.1013849.s001
(PDF)
S2 Table. A.
Peptidoglycan cross-linking analysis; B. Analysis of peptidoglycan containing di-D-Alanine cross-links.
https://doi.org/10.1371/journal.ppat.1013849.s002
(PDF)
S1 Fig. Predictive structure and growth kinetics of dacA mutant strain.
https://doi.org/10.1371/journal.ppat.1013849.s003
(TIFF)
S2 Fig. Bacterial load and serology of animals two-weeks post infection.
https://doi.org/10.1371/journal.ppat.1013849.s004
(TIFF)
S3 Fig. Single example of murine Lyme arthritis caused by dacA mutant strain.
https://doi.org/10.1371/journal.ppat.1013849.s005
(TIFF)
S4 Fig. Bacterial load in ankles 12 weeks post-infection.
https://doi.org/10.1371/journal.ppat.1013849.s006
(TIFF)
S5 Fig. Predicted 3-dimensional structure of p83/100 and model of cell envelope.
https://doi.org/10.1371/journal.ppat.1013849.s007
(TIFF)
Acknowledgments
We would like to thank Kristin Rose Jutras and members of the Jutras lab for their comments on this manuscript. We are grateful for the p83/100 mutant strain and anti-serum for PAP studies, provided by Dr. Jon Skare.
References
- 1. Kugeler K, Schwartz A, Delorey M, Mead P, Hinckley A. Estimating the Frequency of Lyme Disease Diagnoses, United States, 2010–2018. Emerging Infect Dis. 2021;27(2):616.
- 2. Bobe JR, Jutras BL, Horn EJ, Embers ME, Bailey A, Moritz RL, et al. Recent Progress in Lyme Disease and Remaining Challenges. Front Med (Lausanne). 2021;8:666554.
- 3. Keller A, Graefen A, Ball M, Matzas M, Boisguerin V, Maixner F, et al. New insights into the Tyrolean Iceman’s origin and phenotype as inferred by whole-genome sequencing. Nat Commun. 2012;3:698. pmid:22426219
- 4. Walter KS, Carpi G, Caccone A, Diuk-Wasser MA. Genomic insights into the ancient spread of Lyme disease across North America. Nat Ecol Evol. 2017;1(10):1569–76. pmid:29185509
- 5. Steere AC, Strle F, Wormser GP, Hu LT, Branda JA, Hovius JWR, et al. Lyme borreliosis. Nat Rev Dis Primers. 2016;2:16090. pmid:27976670
- 6. Rebman AW, Aucott JN. Post-treatment Lyme Disease as a Model for Persistent Symptoms in Lyme Disease. Front Med (Lausanne). 2020;7:57. pmid:32161761
- 7. Steere AC, Schoen RT, Taylor E. The clinical evolution of Lyme arthritis. Ann Intern Med. 1987;107(5):725–31. pmid:3662285
- 8. Lochhead RB, Strle K, Arvikar SL, Weis JJ, Steere AC. Lyme arthritis: linking infection, inflammation and autoimmunity. Nat Rev Rheumatol. 2021;17(8):449–61. pmid:34226730
- 9. Radolf JD, Caimano MJ, Stevenson B, Hu LT. Of ticks, mice and men: understanding the dual-host lifestyle of Lyme disease spirochaetes. Nat Rev Microbiol. 2012;10(2):87–99. pmid:22230951
- 10. Fraser CM, Casjens S, Huang WM, Sutton GG, Clayton R, Lathigra R, et al. Genomic sequence of a Lyme disease spirochaete, Borrelia burgdorferi. Nature. 1997;390(6660):580–6. pmid:9403685
- 11. Casjens SR, Mongodin EF, Qiu W-G, Dunn JJ, Luft BJ, Fraser-Liggett CM, et al. Whole-genome sequences of two Borrelia afzelii and two Borrelia garinii Lyme disease agent isolates. J Bacteriol. 2011;193(24):6995–6. pmid:22123755
- 12. Lemieux JE, Huang W, Hill N, Cerar T, Freimark L, Hernandez S, et al. Whole genome sequencing of human Borrelia burgdorferi isolates reveals linked blocks of accessory genome elements located on plasmids and associated with human dissemination. PLoS Pathog. 2023;19(8):e1011243. pmid:37651316
- 13. Jutras BL, Lochhead RB, Kloos ZA, Biboy J, Strle K, Booth CJ, et al. Borrelia burgdorferi peptidoglycan is a persistent antigen in patients with Lyme arthritis. Proc Natl Acad Sci U S A. 2019;116(27):13498–507. pmid:31209025
- 14. McClune ME, Ebohon O, Dressler JM, Davis MM, Tupik JD, Lochhead RB, et al. The peptidoglycan of Borrelia burgdorferi can persist in discrete tissues and cause systemic responses consistent with chronic illness. Sci Transl Med. 2025;17(795):eadr2955. pmid:40267217
- 15. Davis MM, Brock AM, DeHart TG, Boribong BP, Lee K, McClune ME, et al. The peptidoglycan-associated protein NapA plays an important role in the envelope integrity and in the pathogenesis of the lyme disease spirochete. PLoS Pathog. 2021;17(5):e1009546. pmid:33984073
- 16. Irazoki O, Hernandez SB, Cava F. Peptidoglycan Muropeptides: Release, Perception, and Functions as Signaling Molecules. Front Microbiol. 2019;10:500. pmid:30984120
- 17. Wolf AJ, Underhill DM. Peptidoglycan recognition by the innate immune system. Nat Rev Immunol. 2018;18(4):243–54. pmid:29292393
- 18. Chan JM, Dillard JP. Attention Seeker: Production, Modification, and Release of Inflammatory Peptidoglycan Fragments in Neisseria Species. J Bacteriol. 2017;199(20):e00354-17. pmid:28674065
- 19. Beck G, Benach JL, Habicht GS. Isolation, preliminary chemical characterization, and biological activity of Borrelia burgdorferi peptidoglycan. Biochem Biophys Res Commun. 1990;167(1):89–95. pmid:2310405
- 20. DeHart TG, Kushelman MR, Hildreth SB, Helm RF, Jutras BL. The unusual cell wall of the Lyme disease spirochaete Borrelia burgdorferi is shaped by a tick sugar. Nat Microbiol. 2021;6(12):1583–92. pmid:34819646
- 21. Vollmer W, Blanot D, de Pedro MA. Peptidoglycan structure and architecture. FEMS Microbiol Rev. 2008;32(2):149–67. pmid:18194336
- 22. Pinho MG, Filipe SR, de Lencastre H, Tomasz A. Complementation of the essential peptidoglycan transpeptidase function of penicillin-binding protein 2 (PBP2) by the drug resistance protein PBP2A in Staphylococcus aureus. J Bacteriol. 2001;183(22):6525–31. pmid:11673420
- 23. Sassine J, Sousa J, Lalk M, Daniel RA, Vollmer W. Cell morphology maintenance in Bacillus subtilis through balanced peptidoglycan synthesis and hydrolysis. Sci Rep. 2020;10(1):17910. pmid:33087775
- 24. Lin T, Gao L, Zhang C, Odeh E, Jacobs MB, Coutte L, et al. Analysis of an ordered, comprehensive STM mutant library in infectious Borrelia burgdorferi: insights into the genes required for mouse infectivity. PLoS One. 2012;7(10):e47532. pmid:23133514
- 25. Sauvage E, Kerff F, Terrak M, Ayala JA, Charlier P. The penicillin-binding proteins: structure and role in peptidoglycan biosynthesis. FEMS Microbiol Rev. 2008;32(2):234–58. pmid:18266856
- 26. Samuels DS, Mach KE, Garon CF. Genetic transformation of the Lyme disease agent Borrelia burgdorferi with coumarin-resistant gyrB. J Bacteriol. 1994;176(19):6045–9. pmid:7928965
- 27. Norris SJ, Howell JK, Garza SA, Ferdows MS, Barbour AG. High- and low-infectivity phenotypes of clonal populations of in vitro-cultured Borrelia burgdorferi. Infect Immun. 1995;63(6):2206–12. pmid:7768600
- 28. Katona LI, Tokarz R, Kuhlow CJ, Benach J, Benach JL. The fur homologue in Borrelia burgdorferi. J Bacteriol. 2004;186(19):6443–56. pmid:15375125
- 29. Miller CL, Karna SLR, Seshu J. Borrelia host adaptation Regulator (BadR) regulates rpoS to modulate host adaptation and virulence factors in Borrelia burgdorferi. Mol Microbiol. 2013;88(1):105–24. pmid:23387366
- 30. Zamba-Campero M, Soliman D, Yu H, Lasseter AG, Chang Y-Y, Silberman JL, et al. Broadly conserved FlgV controls flagellar assembly and Borrelia burgdorferi dissemination in mice. Nat Commun. 2024;15(1):10417. pmid:39614093
- 31. Navarro PP, Vettiger A, Ananda VY, Llopis PM, Allolio C, Bernhardt TG, et al. Cell wall synthesis and remodelling dynamics determine division site architecture and cell shape in Escherichia coli. Nat Microbiol. 2022;7(10):1621–34. pmid:36097171
- 32. Truong TT, Vettiger A, Bernhardt TG. Cell division is antagonized by the activity of peptidoglycan endopeptidases that promote cell elongation. Mol Microbiol. 2020;114(6):966–78. pmid:32866331
- 33. Cho H, Uehara T, Bernhardt TG. Beta-lactam antibiotics induce a lethal malfunctioning of the bacterial cell wall synthesis machinery. Cell. 2014;159(6):1300–11. pmid:25480295
- 34. Lund VA, Wacnik K, Turner RD, Cotterell BE, Walther CG, Fenn SJ, et al. Molecular coordination of Staphylococcus aureus cell division. Elife. 2018;7:e32057. pmid:29465397
- 35. Hernández SB, Dörr T, Waldor MK, Cava F. Modulation of Peptidoglycan Synthesis by Recycled Cell Wall Tetrapeptides. Cell Rep. 2020;31(4):107578. pmid:32348759
- 36. Wacnik K, Rao VA, Chen X, Lafage L, Pazos M, Booth S, et al. Penicillin-Binding Protein 1 (PBP1) of Staphylococcus aureus Has Multiple Essential Functions in Cell Division. mBio. 2022;13(4):e0066922. pmid:35703435
- 37. Gray AN, Egan AJF, Van’t Veer IL, Verheul J, Colavin A, Koumoutsi A, et al. Coordination of peptidoglycan synthesis and outer membrane constriction during Escherichia coli cell division. Elife. 2015;4:e07118. pmid:25951518
- 38. Do T, Schaefer K, Santiago AG, Coe KA, Fernandes PB, Kahne D, et al. Staphylococcus aureus cell growth and division are regulated by an amidase that trims peptides from uncrosslinked peptidoglycan. Nat Microbiol. 2020;5(2):291–303. pmid:31932712
- 39. Monteiro JM, Pereira AR, Reichmann NT, Saraiva BM, Fernandes PB, Veiga H, et al. Peptidoglycan synthesis drives an FtsZ-treadmilling-independent step of cytokinesis. Nature. 2018;554(7693):528–32. pmid:29443967
- 40. Bonnet J, Durmort C, Jacq M, Mortier-Barrière I, Campo N, VanNieuwenhze MS, et al. Peptidoglycan O-acetylation is functionally related to cell wall biosynthesis and cell division in Streptococcus pneumoniae. Mol Microbiol. 2017;106(5):832–46. pmid:28960579
- 41. Wheeler R, Turner RD, Bailey RG, Salamaga B, Mesnage S, Mohamad SAS, et al. Bacterial Cell Enlargement Requires Control of Cell Wall Stiffness Mediated by Peptidoglycan Hydrolases. mBio. 2015;6(4):10.1128/mbio.00660-15.
- 42. Jutras BL, Scott M, Parry B, Biboy J, Gray J, Vollmer W, et al. Lyme disease and relapsing fever Borrelia elongate through zones of peptidoglycan synthesis that mark division sites of daughter cells. Proc Natl Acad Sci U S A. 2016;113(33):9162–70. pmid:27506799
- 43. Gabby ME, Bandara A, Outrata LM, Ebohon O, Ahmad SS, Dressler JM, et al. A high-resolution screen identifies a preexisting beta-lactam that specifically treats Lyme disease in mice. Sci Transl Med. 2025;17(795):eadr9091. pmid:40267215
- 44. Motaleb MA, Corum L, Bono JL, Elias AF, Rosa P, Samuels DS, et al. Borrelia burgdorferi periplasmic flagella have both skeletal and motility functions. Proc Natl Acad Sci U S A. 2000;97(20):10899–904. pmid:10995478
- 45. Charon NW, Goldstein SF, Marko M, Hsieh C, Gebhardt LL, Motaleb MA, et al. The flat-ribbon configuration of the periplasmic flagella of Borrelia burgdorferi and its relationship to motility and morphology. J Bacteriol. 2009;191(2):600–7. pmid:19011030
- 46. Dombrowski C, Kan W, Motaleb MA, Charon NW, Goldstein RE, Wolgemuth CW. The elastic basis for the shape of Borrelia burgdorferi. Biophys J. 2009;96(11):4409–17. pmid:19486665
- 47. Barthold SW, Beck DS, Hansen GM, Terwilliger GA, Moody KD. Lyme borreliosis in selected strains and ages of laboratory mice. J Infect Dis. 1990;162(1):133–8. pmid:2141344
- 48. Barthold SW, Sidman CL, Smith AL. Lyme borreliosis in genetically resistant and susceptible mice with severe combined immunodeficiency. Am J Trop Med Hyg. 1992;47(5):605–13. pmid:1449201
- 49. Kang I, Barthold SW, Persing DH, Bockenstedt LK. T-helper-cell cytokines in the early evolution of murine Lyme arthritis. Infect Immun. 1997;65(8):3107–11. pmid:9234761
- 50. Coburn J, Garcia B, Hu LT, Jewett MW, Kraiczy P, Norris SJ. Lyme Disease Pathogenesis. Curr Issues Mol Biol. 2021;42:473–518.
- 51. Arvikar SL, Steere AC. Diagnosis and treatment of Lyme arthritis. Infect Dis Clin North Am. 2015;29(2):269–80. pmid:25999223
- 52. Park JS, Lee WC, Yeo KJ, Ryu K-S, Kumarasiri M, Hesek D, et al. Mechanism of anchoring of OmpA protein to the cell wall peptidoglycan of the gram-negative bacterial outer membrane. FASEB J. 2012;26(1):219–28. pmid:21965596
- 53. Godlewska R, Wiśniewska K, Pietras Z, Jagusztyn-Krynicka EK. Peptidoglycan-associated lipoprotein (Pal) of Gram-negative bacteria: function, structure, role in pathogenesis and potential application in immunoprophylaxis. FEMS Microbiol Lett. 2009;298(1):1–11. pmid:19519769
- 54. Kovacs-Simon A, Titball RW, Michell SL. Lipoproteins of bacterial pathogens. Infect Immun. 2011;79(2):548–61. pmid:20974828
- 55. Rössler D, Eiffert H, Jauris-Heipke S, Lehnert G, Preac-Mursic V, Teepe J, et al. Molecular and immunological characterization of the p83/100 protein of various Borrelia burgdorferi sensu lato strains. Med Microbiol Immunol. 1995;184(1):23–32. pmid:8538575
- 56. Hauser U, Lehnert G, Lobentanzer R, Wilske B. Interpretation criteria for standardized Western blots for three European species of Borrelia burgdorferi sensu lato. J Clin Microbiol. 1997;35(6):1433–44. pmid:9163458
- 57. Lefebvre RB, Perng GC, Johnson RC. The 83-kilodalton antigen of Borrelia burgdorferi which stimulates immunoglobulin M (IgM) and IgG responses in infected hosts is expressed by a chromosomal gene. J Clin Microbiol. 1990;28(7):1673–5. pmid:2380391
- 58. Perng GC, LeFebvre RB, Johnson RC. Further characterization of a potent immunogen and the chromosomal gene encoding it in the Lyme disease agent, Borrelia burgdorferi. Infect Immun. 1991;59(6):2070–4. pmid:2037367
- 59. Wager B, Shaw DK, Groshong AM, Blevins JS, Skare JT. BB0744 Affects Tissue Tropism and Spatial Distribution of Borrelia burgdorferi. Infect Immun. 2015;83(9):3693–703. pmid:26150534
- 60. McCausland JW, Kloos ZA, Irnov I, Sonnert ND, Zhou J, Putnik R, et al. Bacterial and host enzymes modulate the pro-inflammatory response elicited by the peptidoglycan of Lyme disease agent Borrelia burgdorferi. PLoS Pathog. 2025;21(7):e1013324. pmid:40623106
- 61. Ghosh AS, Young KD. Sequences near the active site in chimeric penicillin binding proteins 5 and 6 affect uniform morphology of Escherichia coli. J Bacteriol. 2003;185(7):2178–86. pmid:12644487
- 62. Nelson DE, Young KD. Penicillin binding protein 5 affects cell diameter, contour, and morphology of Escherichia coli. J Bacteriol. 2000;182(6):1714–21. pmid:10692378
- 63. Matsuhashi M, Wachi M, Ishino F. Machinery for cell growth and division: penicillin-binding proteins and other proteins. Res Microbiol. 1990;141(1):89–103. pmid:2114032
- 64. Morè N, Martorana AM, Biboy J, Otten C, Winkle M, Serrano CKG, et al. Peptidoglycan Remodeling Enables Escherichia coli To Survive Severe Outer Membrane Assembly Defect. mBio. 2019;10(1):e02729-18. pmid:30723128
- 65. Morlot C, Pernot L, Le Gouellec A, Di Guilmi AM, Vernet T, Dideberg O, et al. Crystal structure of a peptidoglycan synthesis regulatory factor (PBP3) from Streptococcus pneumoniae. J Biol Chem. 2005;280(16):15984–91. pmid:15596446
- 66. Wyke AW, Ward JB, Hayes MV, Curtis NA. A role in vivo for penicillin-binding protein-4 of Staphylococcus aureus. Eur J Biochem. 1981;119(2):389–93. pmid:7308191
- 67. Choi U, Park SH, Lee HB, Son JE, Lee C-R. Coordinated and Distinct Roles of Peptidoglycan Carboxypeptidases DacC and DacA in Cell Growth and Shape Maintenance under Stress Conditions. Microbiol Spectr. 2023;11(3):e0001423. pmid:37098975
- 68. Wolgemuth CW. Flagellar motility of the pathogenic spirochetes. Semin Cell Dev Biol. 2015;46:104–12. pmid:26481969
- 69. Sycuro LK, Pincus Z, Gutierrez KD, Biboy J, Stern CA, Vollmer W, et al. Peptidoglycan crosslinking relaxation promotes Helicobacter pylori’s helical shape and stomach colonization. Cell. 2010;141(5):822–33. pmid:20510929
- 70. Sycuro LK, Wyckoff TJ, Biboy J, Born P, Pincus Z, Vollmer W, et al. Multiple peptidoglycan modification networks modulate Helicobacter pylori’s cell shape, motility, and colonization potential. PLoS Pathog. 2012;8(3):e1002603. pmid:22457625
- 71. Frirdich E, Vermeulen J, Biboy J, Soares F, Taveirne ME, Johnson JG, et al. Peptidoglycan LD-carboxypeptidase Pgp2 influences Campylobacter jejuni helical cell shape and pathogenic properties and provides the substrate for the DL-carboxypeptidase Pgp1. J Biol Chem. 2014;289(12):8007–18. pmid:24394413
- 72. Hunashal Y, Kumar GS, Choy MS, D’Andréa ÉD, Da Silva Santiago A, Schoenle MV, et al. Molecular basis of β-lactam antibiotic resistance of ESKAPE bacterium E. faecium Penicillin Binding Protein PBP5. Nat Commun. 2023;14(1):4268. pmid:37460557
- 73. Rybkine T, Mainardi JL, Sougakoff W, Collatz E, Gutmann L. Penicillin-binding protein 5 sequence alterations in clinical isolates of Enterococcus faecium with different levels of beta-lactam resistance. J Infect Dis. 1998;178(1):159–63. pmid:9652435
- 74. Park SH, Choi U, Ryu S-H, Lee HB, Lee J-W, Lee C-R. Divergent Effects of Peptidoglycan Carboxypeptidase DacA on Intrinsic β-Lactam and Vancomycin Resistance. Microbiol Spectr. 2022;10(4):e0173422. pmid:35758683
- 75. Sütterlin L, Edoo Z, Hugonnet JE, Mainardi JL, Arthur M. Peptidoglycan cross-linking activity of l,d-transpeptidases from Clostridium difficile and inactivation of these enzymes by β-lactams. Antimicrob Agents Chemother. 2018;62(1).
- 76. Abdullah MR, Gutiérrez-Fernández J, Pribyl T, Gisch N, Saleh M, Rohde M, et al. Structure of the pneumococcal l,d-carboxypeptidase DacB and pathophysiological effects of disabled cell wall hydrolases DacA and DacB. Mol Microbiol. 2014;93(6):1183–206. pmid:25060741
- 77. Masters EA, de Mesy Bentley KL, Gill AL, Hao SP, Galloway CA, Salminen AT. Identification of Penicillin Binding Protein 4 (PBP4) as a critical factor for Staphylococcus aureus bone invasion during osteomyelitis in mice. PLOS Pathogens. 2020;16(10):e1008988.
- 78. Yang X, Promnares K, Qin J, He M, Shroder DY, Kariu T, et al. Characterization of multiprotein complexes of the Borrelia burgdorferi outer membrane vesicles. J Proteome Res. 2011;10(10):4556–66. pmid:21875077
- 79. Ratet G, Santecchia I, Fanton d’Andon M, Vernel-Pauillac F, Wheeler R, Lenormand P, et al. LipL21 lipoprotein binding to peptidoglycan enables Leptospira interrogans to escape NOD1 and NOD2 recognition. PLoS Pathog. 2017;13(12):e1006725. pmid:29211798
- 80. Braun V. Covalent lipoprotein from the outer membrane of Escherichia coli. Biochim Biophys Acta. 1975;415(3):335–77. pmid:52377
- 81. Magnet S, Bellais S, Dubost L, Fourgeaud M, Mainardi J-L, Petit-Frère S, et al. Identification of the L,D-transpeptidases responsible for attachment of the Braun lipoprotein to Escherichia coli peptidoglycan. J Bacteriol. 2007;189(10):3927–31. pmid:17369299
- 82. Rogers EA, Marconi RT. Delineation of species-specific binding properties of the CspZ protein (BBH06) of Lyme disease spirochetes: evidence for new contributions to the pathogenesis of Borrelia spp. Infect Immun. 2007;75(11):5272–81. pmid:17846117
- 83. Rogers EA, Abdunnur SV, McDowell JV, Marconi RT. Comparative analysis of the properties and ligand binding characteristics of CspZ, a factor H binding protein, derived from Borrelia burgdorferi isolates of human origin. Infect Immun. 2009;77(10):4396–405. pmid:19620346
- 84. Codolo G, Bossi F, Durigutto P, Bella CD, Fischetti F, Amedei A, et al. Orchestration of inflammation and adaptive immunity in Borrelia burgdorferi-induced arthritis by neutrophil-activating protein A. Arthritis Rheum. 2013;65(5):1232–42. pmid:23371320
- 85. Chen Y, Vargas SM, Smith TC 2nd, Karna SLR, MacMackin Ingle T, Wozniak KL, et al. Borrelia peptidoglycan interacting Protein (BpiP) contributes to the fitness of Borrelia burgdorferi against host-derived factors and influences virulence in mouse models of Lyme disease. PLoS Pathog. 2021;17(4):e1009535. pmid:33882111
- 86. Zückert WR. Laboratory maintenance of Borrelia burgdorferi. Curr Protoc Microbiol. 2007;Chapter 12:Unit 12C.1. pmid:18770608
- 87. Jutras BL, Chenail AM, Stevenson B. Changes in bacterial growth rate govern expression of the Borrelia burgdorferi OspC and Erp infection-associated surface proteins. J Bacteriol. 2013;195(4):757–64. pmid:23222718
- 88. Paintdakhi A, Parry B, Campos M, Irnov I, Elf J, Surovtsev I, et al. Oufti: an integrated software package for high-accuracy, high-throughput quantitative microscopy analysis. Mol Microbiol. 2016;99(4):767–77. pmid:26538279
- 89. Barthold SW, Moody KD, Terwilliger GA, Jacoby RO, Steere AC. An animal model for Lyme arthritis. Ann N Y Acad Sci. 1988;539:264–73. pmid:3263827
- 90. Tang X, Cao Y, Booth CJ, Arora G, Cui Y, Matias J, et al. Adiponectin in the mammalian host influences ticks’ acquisition of the Lyme disease pathogen Borrelia. PLoS Biol. 2023;21(10):e3002331. pmid:37862360
- 91. Adusumilli S, Booth CJ, Anguita J, Fikrig E. Passage through Ixodes scapularis ticks enhances the virulence of a weakly pathogenic isolate of Borrelia burgdorferi. Infect Immun. 2010;78(1):138–44. pmid:19822652
- 92. Liu N, Belperron AA, Booth CJ, Bockenstedt LK. The caspase 1 inflammasome is not required for control of murine Lyme borreliosis. Infect Immun. 2009;77(8):3320–7. pmid:19487481
- 93. Morrison TB, Ma Y, Weis JH, Weis JJ. Rapid and sensitive quantification of Borrelia burgdorferi-infected mouse tissues by continuous fluorescent monitoring of PCR. J Clin Microbiol. 1999;37(4):987–92. pmid:10074514
- 94. Jutras BL, Bowman A, Brissette CA, Adams CA, Verma A, Chenail AM, et al. EbfC (YbaB) is a new type of bacterial nucleoid-associated protein and a global regulator of gene expression in the Lyme disease spirochete. J Bacteriol. 2012;194(13):3395–406. pmid:22544270
- 95. Abramson J, Adler J, Dunger J, Evans R, Green T, Pritzel A, et al. Accurate structure prediction of biomolecular interactions with AlphaFold 3. Nature. 2024;630(8016):493–500. pmid:38718835