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

The Peptidoglycan (PG) cell wall of the Lyme disease (LD) spirochete, Borrelia burgdorferi (Bb), contributes to structural and morphological integrity of Bb; is a persistent antigen in LD patients; and has a unique pentapeptide with L-Ornithine as the third amino acid that cross-links its glycan polymers. A borrelial homolog (BB_0167) interacted specifically with borrelilal PG via its peptidoglycan interacting motif (MHELSEKRARAIGNYL); was localized to the protoplasmic cylinder of Bb; and was designated as Borrelia peptidoglycan interacting Protein (BpiP). A bpiP mutant displayed no defect under in vitro growth conditions with similar levels of several virulence-related proteins. However, the burden of bpiP mutant in C3H/HeN mice at day 14, 28 and 62 post-infection was significantly lower compared to control strains. No viable bpiP mutant was re-isolated from any tissues at day 62 post-infection although bpiP mutant was able to colonize immunodeficient SCID at day 28 post-infection. Acquisition or transmission of bpiP mutant by Ixodes scapularis larvae or nymphs respectively, from and to mice, was significantly lower compared to control strains. Further analysis of bpiP mutant revealed increased sensitivity to vancomycin, osmotic stress, lysosomal extracts, human antimicrobial peptide cathelicidin-LL37, complement-dependent killing in the presence of day 14 post-infection mouse serum and increased internalization of CFSC-labeled bpiP mutant by macrophages and dendritic cells compared to control strains. These studies demonstrate the importance of accessory protein/s involved in sustaining integrity of PG and cell envelope during different phases of Bb infection.

Introduction Lyme disease is the most common tick-borne infectious disease in the US with more than 40,000 confirmed and around 300,000 estimated infections occurring each year according to Centers for Disease Control and Prevention [1]. The causative agent of Lyme disease, Borrelia burgdorferi (Bb), is a spirochetal pathogen that is transmitted to humans and other vertebrate hosts following the bite of an infected Ixodes scapularis tick. Bb has a compact genome with limited metabolic capabilities [2][3][4]. As an extreme auxotroph, Bb is constrained by 1) environmental signals; 2) limited/variable levels of key nutrients; and 3) effects of numerous antimicrobial factors impacting its survival in ticks and vertebrate hosts [2,5]. To overcome these bottlenecks, Bb undergoes extensive remodeling of its cell envelope-comprising of an inner cytoplasmic membrane, the peptidoglycan (PG) cell wall and the outer membrane with a constellation of primarily lipoproteins-connecting its metabolism and survival strategies to availability/transport of nutrients to survive and colonize highly divergent hosts. The molecular mechanisms and key determinants that contribute to remodeling and structural integrity of the borrelial cell envelope facilitating host-specific adaptation of Bb are unclear and are open areas of investigation.
A large body of information is available on the metabolic processes and gene regulatory networks involved in the expression of the surface-exposed proteins that enable interactions of Bb with host cells [6][7][8]. Moreover, these interactions also modulate host-response to Bb infection and in the pathogenic mechanisms leading to colonization of Bb in reservoir and dead-end hosts [9]. The role of a number of transporters localized in the periplasmic space and the inner membrane connecting the external environmental milieu to metabolic processes has underscored the importance of changes in the inner membrane impacting assimilation of nutrients that are abundant or sparse during different stages of Bb life cycle [10][11][12][13][14]. Although Bb is a diderm, the absence of lipopolysaccharide and prevalence of a large proportion of phosphatidylcholine in place of phosphatidylethanolamine in the phospholipid bilayer of the inner and outer membranes distinguish the cell envelope of Bb from that of other Gram-negative organisms such as Escherichia coli [15,16]. In addition, the presence of flagella in the periplasmic space confers motility and spiral structure of Bb [17]. It is likely that the peptidoglycan (PG) cell wall confers integrity and modulates host response during the tick and mammalian phases of infection [6].
The biogenesis and synchronization of PG synthesis with cell division and growth rates of Bb under different host-specific conditions are beginning to be understood in greater detail [18,19]. The PG biosynthetic pathway starts with the influx of Short-Chain Fatty Acids (SCFAs) such as acetate that leads to the formation of key substrates such as acetyl CoA, mevalonate and lipid II critical for the biogenesis of PG [19][20][21][22][23][24]. Deletion of genes encoding acetate kinase (ackA) or phosphate acetyl transferase (pta) was lethal for Bb. However, mutants lacking these genes could be rescued with supplementation of mevalonate in the growth medium [23]. Recently, it has been proposed that the breakdown products of borrelial PG accumulate in the medium without being recycled and serves as a stable stimulatory component in the synovial fluid of Lyme disease patients resulting in a sustained pro-inflammatory response [25]. Therefore, a greater understanding of underlying mechanisms that contribute to the biogenesis and integrity of borrelial cell wall will advance strategies to disrupt Bb survival during its infectious cycle.
Previous reports have shown that spirochetes, including B. burgdorferi and Treponemes but not Leptospira spp., have unique PG with a pentapeptide comprising of L-Ornithine (along with L-and D-Alanine, D-Glutamic acid, and Glycine) that cross-links N-acetyl glucosamine (GlcNAc) and N-acetyl muramic acid (MurNAc) polymers [25][26][27][28]. The significance of L-Ornithine as a key structural difference in the PG of Bb in comparison to the presence of meso-diaminopimeleic acid (meso-DAP) and L-Lysine as the branched chain residues in the pentapeptide of many Gram-negative and Gram-positive bacteria, respectively, provides an avenue to dissect the molecular basis of the contribution of borrelial PG to innate and adaptive immune responses against Lyme spirochetes during different stages of infectious cycle [29].
Bioinformatic analysis of Bb genome revealed the presence of a homolog (BB_0167) with similarity to the C-terminal domain of OmpA-C-like proteins and with structural features suggesting its interactions with the pentapeptide of Bb PG [30]. OmpA-C-like homologs are prevalent in many prokaryotes and contribute to the integrity of the cell wall. The C-terminal domain of a periplasmic protein of Neisseria meningitidis, RmpM, shares similarities with the C-terminal domain of OmpA of Escherichia coli with a PG-binding function [31]. Additionally, there are examples of proteins such as Pal, a lipoprotein, and MotB, a flagellar motor protein, known to be associated with the PG. While Pal is anchored to the outer membrane of N. meningitidis via its N-terminal lipid group, the MotB protein has been shown to anchor the inner membrane to PG to immobilize the stator ring of the bacterial flagellar motor [32][33][34]. Moreover, Vibrio cholerae flagellar motor proteins, PomB and MotY, interact with inner membrane and PG [35]. Based on these aforementioned observations and in conjunction with the conserved sequence and structural features, it is feasible that BB_0167 confers stability to the cell wall in particular or enable the cell envelope to withstand a variety of stressors encountered during the infectious cycle of Bb. Moreover, bb_0167 was up-regulated in fed ticks suggesting the relevance of this protein during the adaptation/survival of Bb during different stages of its infectious cycle [36]. The complex and unique organization of cell envelope, a distinct pentapeptide with L-Ornithine as the third amino acid cross-linking the glycan polymers, and the ability for extensive remodeling in different hosts indicate novel structure/function attributes for proteins such as BB_0167 predicted to interact with PG in the survival of Bb within highly disparate host.
In this study, we focused on the biochemical, genetic and phenotypic analysis of a mutant lacking BB_0167, which we designated as Borrelia peptidoglycan interacting Protein (BpiP) based on several lines of evidence and will be referred to as such for the rest of the manuscript.
We localized BpiP to the protoplasmic cylinder/periplasmic fraction of Bb and determined its ability to bind borrelilal PG under in vitro conditions. Moreover, a mutant lacking bpiP displayed a significant colonization deficit in immunocompetent C3H/HeN and BALB/c mice at day 28 and 62 but not in the immunodeficient SCID mice. Rates of acquisition and transmission of bpiP mutant by I. scapularis larvae and nymphs respectively, from and to C3H/HeN mice, were significantly lower than the control strains reflecting the lower burden of bpiP mutant acquired from mammalian host. Phenotypic analysis of the bpiP mutant under in vitro conditions indicated increased sensitivity to select antibiotics, osmotic stress, lysosomal extracts, select antimicrobial peptides, increased phagocytosis by dendritic cells and activated macrophages and susceptibility to complement-dependent killing in the presence of serum from immunocompetent mice obtained at day 14 post-infection. Taken together, our data suggest that BpiP plays a role in facilitating Bb to withstand environmental stressors and soluble/ cellular mediators of host immune response. These observations open avenues to exploit the novel structural and functional features of proteins that contribute to the stability of the cell wall/cell envelope to help advance strategies to limit the survival of Bb during the tick and mammalian phases of infection.

Bioinformatic analysis of BpiP
Sequence analysis revealed that BpiP is a~42 kDA (381 amino acids) protein with homology to the C-terminal region of OmpA-C-like proteins with a conserved PG interaction domain spanning the residues 315 and 338. It is predicted that this domain is involved in the formation of non-covalent interactions between BpiP and PG of Bb (Fig 1A). The gene encoding for BpiP is located in between malQ and dksA homologs of Bb within the linear chromosome and is a distinct transcriptional unit [2]. Recently, the crystal structure of BpiP was solved and compared with the C-terminal domains of OmpA-C like homologs from Salmonella enterica serovar typhimurium (St) and Acenitobacter baumanii (Ab) (Fig 1B, adapted and modified from [30]). Key findings relevant to the current study include the prediction that a Glutamic acid (E315) residue in BpiP potentially interacts with L-Ornithine in the Bb pentapeptide in place of Aspartic acid present in other OmpA-C homologs that facilitate this interaction with meso-DAP present in the pentapeptide of St or Ab [30]. Additionally, the crystal structure of BpiP revealed the binding of sulfate anion (SO 4 2-, assimilated from the crystallization buffer containing ammonium sulfate) outside of the PG binding domain interacts with a conserved residue at position R330 reflecting the bidendate salt bridges and hydrogen bond formation contributing to its interactions with PG [30]. These observations suggest the BpiP has structural features to interact with L-Ornithine that is unique to the pentapeptide crosslinks present in borrelial PG.

Localization of BpiP in Bb
Since BpiP was annotated as an OmpA/C like protein, we determined the localization of BpiP by fractionating Bb into Outer-membrane vesicles (OMVs) and protoplasmic cylinder (PC) fractions. The periplasmic or spheroplasmic proteins of Bb were also analyzed for presence of BpiP. As shown in Fig 2B, immunoblot analysis revealed presence of BpiP in the PC fraction while a known outer membrane protein (P66) localized to the OMV fraction. Moreover, PotA, a known inner membrane protein, also localized to the PC fraction as expected. The total protein profiles of the OMV and PC fractions are shown in Fig 2A. The localization of the BpiP in the PC fraction was also determined by transforming bpiP mutant with a borrelial shuttle vector expressing BpiP under the control of its native promoter and fused to 3XFLAG-tag ( Fig  2E). Immunoblot analysis using anti-FLAG-tag antibodies revealed the localization of BpiP fused to FLAG-tag in the PC fraction while P66 and PotA localized to the OMV and PC fractions respectively, as expected ( Fig 2D). Protein profiles of the OMV and PC fractions of bpiP mt carrying pYC153 are shown in Fig 2C. Immunoblot analysis of Bb (bpiP mutant carrying pYC153), spheroplasmic and periplasmic proteins using anti-FLAG-Tag antibodies revealed the presence of BpiP in the periplast fraction although the amount of BpiP within the whole cells and spheroplasmic fractions were higher (Fig 2F, α-FLAG-Tag). As a control, FlaB was localized abundantly in the periplasmic fraction of Bb consistent with its localization within the periplasmic space of Bb (Fig 2F, α-FlaB). However, there was no detection of gfp in the periplast fraction suggesting that there is no contamination of the cytosolic proteins under the conditions used to prepare the periplasts and spheroplasts (Fig 2F, anti-gfp). These observations suggest that although BpiP (BB0167) was annotated as an Outer Membrane Protein A/C like protein, it is localized in the PC fraction and is likely to accumulate or be detected in the periplasm of Bb.

BpiP binds to borrelial PG
Recombinant BpiP (rBpiP) with N-terminal Maltose Binding Protein (MBP) was purified and MBP was cleaved using TEV protease (S1 Fig). Interactions of BpiP with purified borrelial PG were analyzed after cross-linking with the reagent DTSSP as described in the Materials and Methods (Figs S3 and 3A). BpiP binding to PG was detected using anti BpiP serum generated against C-terminal region of BpiP purified using 7X His-tag [30]. The full-length wild-type BpiP bound borrelial PG (Fig 3B, Lane 7). However, BpiP with a site-specific change at E315A (Fig 3B, Lane 5) or BpiP with 7 conserved residues of the PG-binding motif (shown in red in Fig 1A) replaced with alanines or glycine showed no binding to borrelial PG (Fig 3B, Lane 6). As expected, purified PG by itself did not react with BpiP antibodies (Fig 3B, Lane 4) and the levels of wild type BpiP, E315A and 7SDMs used in the binding studies were similar as shown by their reactivity to anti BpiP antibodies (Fig 3B, Lanes 1-3). This biochemical analysis demonstrates the binding of BpiP to borrelial PG via its PG-binding motif. In addition, we also determined if BpiP interacted with heterologous PG from E. coli and B. subtilis that have meso-DAP or L-Lysine as the third amino acid in their pentapeptide crosslinks, respectively, using the procedure described for BpiP interaction with borrelial PG (Fig 3C). Immunoblot analysis using anti-BpiP serum revealed interaction of BpiP only with borrelial PG (Fig 3D, Lane 6) but not with E. coli or B. subtilis PG (Fig 3D, Lanes 4 and 5) or with any of the purified PGs ( Fig 3D, Lanes 1-3). These observations demonstrated that BpiP interacted with borrelial PG and that the nature of the composition of the pentapeptide crosslink is critical for this interaction.

bpiP is transcriptionally up-regulated under in vitro growth conditions mimicking midgut of fed ticks
Shifting cultures from conditions mimicking midgut of unfed (pH7.6/23˚C) to that of fed ticks (pH6.8/37˚C) has been shown to drive significant alterations in gene expression in response to changes in the environmental signals [37][38][39]. Previous studies have shown that the transcriptional levels of bpiP were elevated in fed infected nymphs [36]. As shown in Fig 4, the transcriptional levels of bpiP were significantly higher in Bb grown under in vitro conditions mimicking the fed ticks (pH6.8/37˚C) compared to that grown under unfed tick condition (pH 7.6/23˚C). It should be pointed out that the in vitro growth conditions mimicking fed Fractionation of (A) Bb wild type strain B31-A3 and (C) bpiP mutant expressing Flag-tagged BpiP, in trans, transformed with E) plasmid pYC153. pYC153 was derived from pTM61 that constitutively express green fluorescent protein (GFP) under the control of the PflaB promoter. The construct also carries determinants for replication in both E. coli and B. burgdorferi, gentamicin resistance (aacC1) cassette, upstream regions of bpiP and bpiP fused to 3x-FLAG tag. Spirochetes were fractionated into outer membrane vesicle (OMV) and protoplasmic cylinder (PC) fractions as described in the Materials and Methods section. Fractions were separated on 12% SDS-PAGE gel stained with Coomassie brilliant blue (A, C) or transferred to PVDF membranes and subjected immunoblot analysis (B, D) using specific antisera indicated to the right of each blot. (F). Periplasmic and spheroplasmic fractions were separated on SDS-PAGE, transferred to PVDF membrane as described in the Materials and Methods section and subjected to immunoblot analysis with anti-FLAG tag, anti-Green Fluorescent Protein (gfp) or anti-FlaB antibodies. Blots were developed using goat-anti-mouse or anti-rabbit IgG conjugated with HRPO as secondary antibody and Enhanced Chemiluminescence. Lanes 1) Outer membrane vesicle fraction (OMV) 2) Periplasmic cylinder (PC). Note detection of BpiP or BpiP-FLAG only in the PC fraction using either anti-BpiP or anti-FLAG antibody. P66 localized to OMV and PotA to PC fraction in both wild type and bpiP mutant complemented with pYC153. Molecular masses in kDa are indicated to the left. https://doi.org/10.1371/journal.ppat.1009535.g002

PLOS PATHOGENS
ticks do not completely reflect the in vivo conditions but nonetheless increased transcriptional levels of bpiP under fed-tick conditions is suggestive of the key environmental cues that alter its levels of expression.

Generation and in vitro growth phenotype of bpiP mutant and complemented strains
In order to determine the role of BpiP in the patho-physiology of Bb, we undertook a genetic approach by deleting bpiP in Bb strain B31-A3 that has the full complement of infection-associated plasmids [40][41][42][43]. As shown schematically in Fig 5A, we replaced bpiP with streptomycin resistance gene (aadA) under the control of a borrelial promoter P flgB and selected mutants in the presence of streptomycin (50 μg/ml). A cis-complemented strain was generated by restoring an intact copy of bpiP with gentamicin resistance gene (GentR) under the control of P flgB (Fig 5A, ct) and selected in the presence of gentamicin (50 μg/ml). The deletion and ciscomplemented strains were confirmed by PCR using primers flanking bpiP as shown schematically. There was no transcription of bpiP in the mutant strain compared wt or ct strains while the levels of flaB were not statistically different between all three strains ( Fig 5B). The total protein lysates from each of these strains resolved by SDS-12.5% PAGE and stained with Coomassie Blue served as a loading control (Fig 5C). Antiserum generated against antibiotic resistance proteins (recombinant streptomycin-resistance and gentamicin-resistance proteins) was used   Total RNA from wt, mt and ct grown at pH7.6/32˚C was converted to cDNA and quantitative real-time PCR was performed as described in the Materials and Methods section. The levels of bpiP transcripts were absent as expected in the mt compared the control strains. The change in the C t value for each sample was subjected to an unpaired t test in GraphPad Prism software. The asterisk indicates the levels of significance as follows: ��� p<0.001. Total protein lysates from wt (Lane 1), mt (Lane 2) and ct (Lane 3) were separated on 12% SDS-PAGE gel and (C) stained with Coomassie blue or (D) transferred to PVDF membrane for immunoblot analysis using antibodies specific to streptomycin resistance (α-Strep, mt) or gentamicin resistance (α-Gent, ct) proteins synthesized due to the expression of aadA and gentR resistance genes under the control of borrelial promoter P flgB .
https://doi.org/10.1371/journal.ppat.1009535.g005 to determine the synthesis of streptomycin resistance protein in the bpiP mutant and gentamicin resistance protein in the cis-complemented strain, respectively, while the parental strain showed no reactivity to either antisera as expected ( Fig 5D). We also confirmed the deletion and restoration of an intact copy of bpiP by whole genome sequencing of DNA from wt, mt and ct strains [41]. Moreover, sequence analysis of the genomic DNA also revealed that the plasmid profile was identical between wt, mt and ct strains. There were no significant differences in the growth rates of wt, mt and ct strains when propagated under in vitro conditions reflecting growth in laboratory (pH7.6/32˚C); un-fed tick midgut (pH7.6/23˚C); or fed tick midgut (pH6.8/37˚C; S4 Fig). A trans-complemented strain was generated by transforming the bpiP mutant using a borrelial shuttle vector pYC153 obtained by cloning 1kb upstream of bpiP and bpiP-3XFLAG-tag at the Kpn1 site present in the plasmid pTM61 ( Fig 2E) [44]. The trans-complemented strain synthesized BpiP that was localized to the protoplasmic cylinder (PC) fraction, which was detected by immunoblot analysis using anti-FLAG-tag antibodies ( Fig 2D).

Synthesis of select borrelial proteins in bpiP mutant
In order to determine if the deletion of bpiP resulted in changes in key borrelial proteins, we analyzed whole cell protein lysates of wt, mt and ct strains propagated under in vitro conditions either mimicking the midgut of unfed (  [19,38,40,45]. Levels of PotD and SodA were elevated at the unfed tick condition while levels of several proteins such as P66, ChbA and FlaB were similar under both conditions serving as controls for equal loading of proteins. The levels of the aforementioned borrelial proteins were similar in wt, mt and ct indicating that the deletion of bpiP did not have any significant impact on the levels of several key pathogenesis-related lipoproteins and select enzymes involved in metabolic and oxidative stress response pathways.

bpiP mutant is attenuated for infection in C3H/HeN mice
We challenged C3H/HeN mice with 1x10 5 wt, mt or ct strains via needle inoculation and determined the levels of infectivity at day 14, 28 (n = 3; three independent experiments) or 62 (n = 3, two independent experiments) post-infection and select tissues were isolated aseptically as indicated in Table 1. Isolated tissues were inoculated into BSKII medium and scored for growth after a single blind pass at day 5 by examining cultures using dark field microscopy at day 15 post-isolation in primary and blind-pass cultures. As shown in Table 1, the mt strain was isolated from different tissues in 8 out of 9 mice at day 14 while only 4 out of the 9 mice supported growth of the mt at day 28. But at day 62, none of the C3H/HeN mice (6 out of 6) were infected with the mt strain demonstrating a significant survival defect during long-term infection. This is in contrast to all mice being infected with wt and ct strains at day 14, 28 and 62 post-infection. We further enumerated the number of spirochetes in select tissues (skin, one lymph node, spleen and one joint) by quantitative real time PCR (Fig 7). While it was possible to recover mt spirochetes from infected tissues in 8 out 9 mice at day 14, qPCR analysis revealed a significant reduction in the Bb burden in all tissues examined such as skin, lymph node, spleen and one of the joints compared to the wt and ct strains (Fig 7A). A significant reduction in Bb burden in mice infected with mt at day 28 in the tissues examined compared to the control strains ( Fig 7B) while tissues from all mice infected with mt did not exhibit detectable pathogen burden by qPCR at day 62 ( Fig 7C). These infectivity studies clearly demonstrated that BpiP plays a role in the survival of Bb in C3H/HeN mouse model of Lyme disease.

Infectivity analysis of bpiP mutant in immunodeficient SCID mice
To determine if the induction of the adaptive immune response of the host contributes to the reduced infectivity of mt spirochetes at day 28 and 62, we compared the levels of infection of the mt in immunocompetent BALB/c and immunodeficient SCID mice. As shown in Table 2, no viable mt spirochetes were detected following infection of BALB/c mice at day 28 while viable spirochetes were isolated from all tissues tested from SCID mice. In addition, as shown in Fig 8, relative spirochetal burden from select tissues by qPCR revealed a significant reduction in Bb burden in tissues from BALB/c mice ( Fig 8A) while the mt spirochetes were readily Total protein lysates of wt, mt, ct strains grown under conditions mimicking midgut of unfed tick (Lane 1, pH 7.6/ 23˚C) and shifted to conditions mimicking midgut of fed tick (Lane 2, pH 6.8/37˚C) were separated on a 12% SDS-PAGE gel, transferred to PVDF membrane and subjected to immunoblot analysis using antisera against various borrelial proteins indicated to the right of each blot. Goat antimouse HRPO-conjugated antibodies were used as secondary antibodies and blots were developed using ECL system. Molecular masses in kilo Daltons (kDa) are indicated to the left. ( � indicate protein samples diluted 1:10 due to quality of the antibody reagent).
https://doi.org/10.1371/journal.ppat.1009535.g006 detectable in tissues from the SCID mice ( Fig 8B). These observations suggest the role of the adaptive immune response of the host contributes to the in vivo phenotype of the bpiP mutant. Table 1. Analysis of infectivity in C3H/HeN mice with wt, bpiP mt and bpiP ct strains at day 14, 28 and 62 post infection.

Strains
No. of cultures positive/total number. Total number infected

bpiP mutant is attenuated for survival in the mouse-tick mouse cycle of Bb infection
We then determined if the attenuated infectivity phenotype observed in the mammalian host following needle challenge could be reproduced using a tick infection model. The mouse-tickmouse cycle of infection was performed as shown schematically in Fig 9A. We challenged naive C3H/HeN mice with wt, mt and ct strains via intradermal needle inoculation with 10 5 spirochetes per mouse and allowed I. scapularis larvae to feed to repletion on infected mice at day 5 post-challenge. Three fed larvae from each mouse (total 9 for each strain) were processed for qPCR to determine the spirochetal burden in fed larvae ( Fig 9B) and three fed larvae from each mouse (total 9 for each strain) were examined for presence of spirochetes ( Fig 9E). The rest of the fed larvae were allowed to molt to nymphs in an incubator as described in Materials and Methods section. As shown in Fig 9B, there was a significant reduction in number of spirochetes in larvae fed on mice infected with mt compared to that from mice infected with wt and ct strains. We then allowed 10 infected nymphs to feed to repletion on naïve C3H/HeN (n = 3, two biological replicates, data from both experiments combined in Table 3 https://doi.org/10.1371/journal.ppat.1009535.t002

Fig 8. Quantitative real-time PCR analysis of spirochetal burden in BALB/c or SCID mice infected with bpiP mutant.
Groups (n = 3) of 4-week old immunocompetent BALB/c or immunodeficient SCID mice were intradermally infected via needle inoculation with wt, mt and ct strains with 10 5 spirochetes per mouse and were analyzed. Total genomic DNA was isolated from tissue (skin, lymph node, spleen and one joint) using the High Pure PCR template preparation kit and was then subjected to quantitative real-time PCR. Numbers of borrelial flaB copies were normalized against total mouse β-actin. Data analyzed using a one-way analysis of variance test with Tukey's multiple-comparison test. Statistical significance was accepted when p values were less than 0.05. Tissues samples analyzed were collected at day 28 from (A) BALB/c (B) SCID mice. https://doi.org/10.1371/journal.ppat.1009535.g008 and examined 3 fed nymphs from each mouse (total 9) infected with wt, mt or ct strains for the presence of spirochetes by qPCR ( Fig 9C) or by dark field microscopy ( Fig 9E). As shown in Fig 9C, there was a significant reduction in the number of spirochetes in fed nymphs infected with mt compared to nymphs infected with wt or ct strains. The levels of transmission of wt, mt or ct to naive C3H/HeN mice following challenge with infected nymphs were determined by isolating different tissues aseptically for isolation of viable spirochetes in BSKII borrelial growth medium (Table 3) or by qPCR for spirochete burden ( Fig 9D) at day 14 post-tick challenge. Mutant spirochetes were isolated from the skin and lymph nodes of 1 out 6 C3H/ Naïve I. scapularis larvae were fed to repletion on C3H/HeN mice infected with wt, mt or ct strains with 10 5 spirochetes per mouse via needle inoculation. Bb burden in a proportion of fed larvae (3 per mouse) determined by (B) qPCR or (E) examined for the presence of spirochetes by dark field microscopy. Another proportion of fed larvae were allowed to molt to nymphs in an incubator maintained at 22˚C with 90% humidity and a 15-hour light /9-hour dark cycle. Groups (n = 3) of 4-week-old naïve C3H/HeN female mice were challenged with 10 infected nymphs and after nymphs fed to repletion. Bb burden was determined from 3 nymphs per mouse by (C) qPCR or (E) examined for the presence of spirochetes in the midgut by dark field microscopy Infectivity analysis was done at day 14 post-infection by propagating several tissues for the presence of spirochetes (Table 3). A portion of spleen, one lymph node, spleen and one joint was used to extract total genomic DNA and (D) Bb burden was determined quantitative real time PCR. The infectivity analysis was done twice and results of both the experiments are reported together for mice infected with nymphs. Numbers of borrelial flaB copies were normalized against total mouse β-actin copies. The differences between the number of wt and mt or mt and ct in mouse tissues were statistically analyzed using a one-way analysis of variance test with Tukey's multiple-comparison test. Statistical significance was accepted when p values were less than 0.05.
https://doi.org/10.1371/journal.ppat.1009535.g009 HeN mice challenged with infected nymphs while there were no viable spirochetes in other tissues examined. This is in contrast to mice challenged with nymphs infected with wt or ct strains where every tissue was positive for viable spirochetes. Moreover, these mice had significantly higher Bb burden compared to mice challenged with nymphs infected with the mutant. These studies demonstrated that bpiP mt was acquired in lower numbers by larvae from infected mice leading to reduced transmission by infected nymphs to naive C3H/HeN mice.

Sensitivity of bpiP mutant to vancomycin
We then investigated the mechanistic basis for lack of infectivity in mouse models of infection and to connect the functional contribution of BpiP via its interactions with borrelial peptidoglycan to infectivity of Bb in mouse/tick models of LD. We performed a series of analyses to determine the role of BpiP in the ability of spirochetes to withstand exogenous or endogenous anti-bacterial effector molecules that specifically cause stress to the bacterial cell wall/cell envelope. Vancomycin, a glycopeptide antibiotic that inhibits PG synthesis by non-covalent binding to D-alanyl-D-alanine (D-ala-D-ala) terminus of the pentapeptide cell wall precursor displays activity against Bb under in vitro and in vivo growth conditions and has been shown to reduce the cell wall stiffness and motility of Bb [46,47] As shown in Fig 10A, 0.5 and 1.0 μg /ml vancomycin resulted in a significant reduction in the growth of mt compared to wt or ct strains. Concentrations lower than 0.5 or greater than 1.0 μg /ml did not result in significant differences in the growth rates between mt and the control strains. Colony forming units (cfus) of cultures with 0 or 0.5 μg/ml of vancomycin at day 5 post-treatment were obtained by plating on BSKII agar overlays. The mt exhibited a significant reduction in the number of cfu/ ml compared to wt or ct strain following treatment with 0.5 μg/ml of vancomycin ( Fig 10B).

Sensitivity of bpiP mutant to increased osmotic stress
bpiP mutant had a lower growth rate in the presence of 200 mM of sodium chloride above the levels normally present in the BSKII borrelial growth medium after 120 hrs of growth ( Fig  11A). The bpiP mutant had significantly lower cfu/ml with 200 mM concentration of sodium chloride compared to wt or mt or when grown with no additional supplementation of NaCl ( Fig 11B). These observations suggest that lack of BpiP decreases the ability of Lyme spirochetes to withstand select physiological stressors.

Sensitivity of bpiP mutant to lysosomal extracts
We proceeded to determine if host-derived stressors that targeted the cell wall/cell envelope would also lead to a readily observable phenotype in the bpiP mutant. Numerous host-derived antibacterial effectors play a role in controlling bacterial infections. Many of the effectors impact the cell envelope and PG of bacterial cells. Since the bpiP had a significant colonization deficit within 14 days, post-infection, in immunocompetent mouse models, we determined the effects of lysosomal extracts from bone-marrow derived dendritic cells from C3H/HeN or C57BL6 mice. As shown in Fig 12A, bpiP mt incubated with lysosomal extract had a significant growth defect compared to the wt and ct strain. Moreover, the bpiP mutant also had significantly lower cfu/ml when exposed to 10% lysosomal extract compared to wt or mt or when grown without lysosomal extract ( Fig 12B).

Sensitivity of bpiP mutant to human or murine antimicrobial peptides, LL37 or mCRAMP
We then tested the phenotype of the mt in the presence of host-derived antimicrobial peptides from rodents and humans. DC lysosomal extract contains many different antimicrobial effectors [48]. Among these, antimicrobial peptides may play a role in reducing the infectivity of bpiP mutant in the mammalian host. Moreover, during skin infections, cathelicidin-type antimicrobial peptides are induced [49]. The cathelicidin α-helical peptides such as LL37 can rapidly permeabilize the cell membrane of bacteria [50]. Thus, we tested the sensitivity of wt, mt and ct strains with 100 μg/ml of human and mouse cathelicidins, LL37 and mCRAMP, respectively, in BSKII medium at pH6.8 /37˚C and observed that mt was sensitive to human cathelicidin ( Fig 13A) while mCRAMP had no inhibitory effect on mt compared to the control strains (S2 Fig). In addition, the mutant had significantly lower cfu/ml in the presence of 100 μg/ml of LL37 compared wt or ct strains while there were no significant differences in the cfu/ml of these strains in the absence of LL37 (Fig 13B). It is possible the effects of these antimicrobial peptides require environments that are more acidic than that conferred with BSKII medium at pH6.8, although lower pH as such has a higher growth inhibitory effect on Lyme spirochetes. These observations suggest that BpiP contributes to the integrity of the cell wall/ cell envelope against multiple antimicrobial effectors that could synergistically increase the sensitivity of bpiP mutant where a combination of these effectors are prevalent such as within a mammalian host.

Increased susceptibility of bpiP mutant to macrophages/dendritic cells
There was a significant reduction in the number of mt spirochetes detected in C3H/HeN mice by qPCR at day 14 post-infection although there were viable spirochetes in select tissues isolated from infected mice (Fig 7A). This significant reduction in bpiP burden indicated the inhibitory role of cellular and soluble mediators of innate immunity that limited the colonization of the mt compared wt or ct strains. We, therefore, used Imaging Flow Cytometry that allows for both qualitative (Fig 14A and 14C) and quantitative measurement (Fig 14B and  14D) of levels of uptake of Bb by macrophage and dendritic cells [41]. Quantitative measurement of uptake of Bb revealed that the bpiP mt has higher percentage of phagocytosis by activated macrophages (F480 + ) and dendritic cells (CD11c + ) than control strains, suggesting that the increased clearance of bpiP mutant during early stages of infection could partly explain the significant reduction in bpiP burden in infected C3H/HeN mice by quantitative real time PCR (Fig 7A) even at day 14 post-infection. These observations demonstrate that lack of bpiP results in increased uptake of mt spirochetes by cellular mediators of innate immunity during the initial stage of infection and likely contribute to reduce infectivity of bpiP mt in immunocompetent mice. Therefore, its survival through tick and mammalian phases of infection is reliant on appropriate levels of expression of multiple determinants conferring resistance to these host-derived factors.

Complement dependent killing of bpiP mutant
The development of the adaptive immune response as a contributing factor in reducing the survival of the bpiP mutant was predicted as a possible mechanism to reduce the bpiP burden quantitatively in immunocompetent mouse models of infection [51]. The level of antibody titers from C3H/HeN mice at day 14, 28 and 62 post-infection was determined by ELISA using whole cell lysates of wild type Bb. Consistent with above hypothesis, as shown in Fig 15A, the levels of antibody titers at day 14 was not statistically different in sera from mice infected with wt, mt or ct strains. However, the serum from mice infected with mt had significantly lower levels of antibodies at day 28 and 62 compared to the control strains consistent with reduced pathogen burden in these mice (Fig 15B and 15C). In order to determine if the levels of antibodies induced within 14 days were borrelicidal, we determined the complement-dependent killing of serum from mt infected mice. As shown in Fig 15D, the survival of mt was significantly lower compared to the control strains in the presence of guinea pig complement and day 14 serum from mice infected with mt. These observations indicate that the bpiP mt is unable to induce long lasting adaptive immune response and is cleared during the early stage of infection within the mammalian host due to complement mediated events.

Attenuation of infectivity of bpiP mutant is not due to lack/loss of lp28-1/ VlsE
The vls locus encoded on the linear plasmid 28-1 has been shown to play a role in the persistence of Bb in reservoir hosts by contributing to antigenic variation and enabling the survival of Bb in the presence of an intact adaptive immune response [52,53]. It is feasible that the loss of infectivity of bpiP mutant was due to lack of lp28-1 and surface expression of VlsE that may impact the ability of the bpiP mutant to colonize immunocompetent C3H/HeN mice. Both lp28-1 and lp25 were readily detected by PCR using plasmid specific primers while a clonal isolate of Bb-B31 strain lacking lp28-1 did not have an amplicon serving as a negative control (Fig 16). Immunoblot analysis of protein lysates from wt, mt and ct had similar levels of VlsE while the clonal isolate lacking lp28-1 did not show any reactivity (Fig 17B). Immunoblot analysis of spirochetes subjected to Proteinase K (PK) treatment revealed similar surface exposure of VlsE ( Fig 17C). There were no changes in the levels of FlaB with or without PK treatment as FlaB is a periplasmic protein inaccessible to Proteinase K. These studies clearly demonstrate that the loss of infectivity of bpiP mutant in immunocompetent mice is not due the lack of lp28-1, lp25, levels and location of VlsE.

Discussion
The Peptidoglycan (PG) cell wall is an essential component of the cell envelope contributing to the survival of bacterial pathogens under a variety of inhospitable environments [54]. Therefore, the biogenesis, composition and accessory determinants that provide structural and functional integrity to cell wall play a critical role during the colonization and persistence of Bb during the tick and mammalian phases of infection [19,23]. Moreover, integration of synthesis and organization of multiple cell envelope/cell wall components with central and intermediate metabolism synchronized with regulatory effects on motility, morphology, and virulence underscore the importance of the role of these determinants in the pathophysiology of Lyme spirochetes. Recent findings that 1) discrete zones of PG are synthesized to elongate Bb; 2) PG is a persistent antigen in Lyme arthritis patients capable of inducing a pro-inflammatory response; and 3) PG fragments accumulate in the medium rather than recycled for generation of new PG have opened avenues to explore the role of pathogen-specific differences associated with cell wall or cell envelope of Bb as unique targets to interfere with Lyme disease pathogenesis [18,25]. A major finding from this study is the role of a borrelial homolog, BpiP, annotated as OmpA-C like protein, in the survival of Bb in mammalian hosts and its role in contributing to the integrity of borrelial PG (Fig 1). Unlike other cannonical OmpA-C protein in E. coli that connects PG to the outer membrane, BpiP is localized to the protoplasmic cylinder suggesting the possibility of anchoring the PG to the inner membrane (Fig 2). In addition, BpiP-3X-FLAG-tag protein was detected with the periplast fraction although BpiP does not have a typical signal sequence to facilitate its secretion to the periplasmic space ( Fig 2F). It should be pointed out the levels of BpiP were lower in the periplast fraction and it is feasible that BpiP is localized in the cytosol or in the inner membrane of Bb for interaction with L-Ornithine and may potentially serve to "trap and transport" this non-proteogenic amino acid following its biogenesis from citrulline due to the enzymatic functions of ornithine carbamoyltransferase (ArgF, BB0842). Alternatively, it is possible for BpiP to anchor the borrelial PG to the inner membrane or potentially sequester L-Ornithine assimilated from the host environment requiring BpiP localization within the periplasmic space. The signal-dependent synthesis (temperature, pH and stage of cell division) and partitioning of BpiP is unknown and additional studies are needed to delineate how BpiP influences various steps involved in the biogenesis of borrelial PG. The major flagellin of Bb, FlaB, also lacks a canonical signal sequence but is secreted to the periplasm via the Type III secretory pathway or by interaction with another periplasmic protein FlaA with a canonical signal sequence that facilitates its secretion into the periplasmic space [17,55]. There are examples of non-flagellar proteins being secreted via the flagellar Type III secretory pathway [56]. It is possible that BpiP is chaperoned by another protein to the periplasmic space or utilizes known or unknown secretory pathways to be localized in the periplasmic space to interact with the pentapeptide of borrelial PG.
Consistent with prior structural analysis that predicted interaction of BpiP with L-Ornithine via a glutamic acid residue at position 315 [30], our cross-linking studies demonstrated that the recombinant BpiP with E315A or with 7 site-specific changes replacing the residues of the peptidoglycan-binding motif was unable to bind to borrelial PG unlike the native BpiP (Fig 3A and 3B). Moreover, BpiP interacted with borrelial PG but not with heterologous PG from E. coli and B. subtilis that had meso-DAP or L-Lysine as the third amino acid reflecting the specificity of BpiP-Bb PG interactions (Fig 3C and 3D). Therefore, BpiP appears to exert a pathogen-specific function consistent with the unique composition of the pentapeptide of Bb PG.
Biogenesis of PG in the absence of amino acid biosynthetic pathway/s for making L-Ornithine suggests the possibility that Bb acquires this non-proteogenic amino acid from the growth media or from host-derived sources or generates this molecule via an alternate pathway [2,5]. Due to the non-proteogenic nature of L-Ornithine, it is unlikely that OppA-dependent peptide transport system is involved in the process of transporting L-Ornithine [12,13]. Prior studies have shown no significant phenotypic differences in the phenotype of Bb following supplementation of L-Ornithine in borrelial growth medium [57]. The presence of a homolog, BB0842, annotated as ornithine carbamoyltransferase (ArgF) suggests possibility of Arginine being converted to Citrulline, which in turn can be converted to L-Ornithine by ArgF [57,58]. Moreover, the presence of ArcA (BB0841, arginine deaminase) and ArgF provides for a source of L-Ornithine for PG biogenesis [28,40,57,[59][60][61]. Alternatively, it is possible that Bb is able to acquire host derived L-Ornithine-a naturally occurring non-proteogenic amino acid obtained via the actions of Arginase on Arginine that is prevalent in different hosts. Similarly, mammalian sources of N-acetyl hexosamine could be harnessed during the biogenesis of borrelial PG.
Sequence analysis of BpiP homologs present in the Relapsing Fever (RF) spirochetes (B. hermsi and B. turicatae) reveal the presence of an alanine at position 315 which raised the issue of levels of interactions between these homologs with L-Ornithine as the third amino acid in the pentapeptide of RF PG. In addition to this difference, RF BpiP homologs also have different amino acids within the PG binding domain that could contribute to the interaction between RF BpiP and its PG (Fig 1A). It is also possible that the RF spirochetes carry L-Lysine as the third amino acid and therefore A 315 may have little impact on the interactions between the RF BpiP and its PG. Additional studies are warranted to explore the significance of the difference between BpiP homologs of LD and RF spirochetes.
In addition to the source of L-Ornithine, bioinformatic analysis of genomes of B. burgdorferi revealed presence of several conserved homologs encoding for enzymes that sequentially add various components of the PG [19,23]. Among these homologs, borrelial MurE (BB0201) is annotated as UDP-N-acetylmuramyl tripeptide synthetase and as UDP-N-acetylmuramoylalanyl D-glutamate-2,6-diaminopimelate ligase while MurE homolog in B. burgdorferi strain CA382 is annotated as UDP-N-acetylmuramoylalanyl D-glutamate-L-ornithine ligase. While there are multiple annotations of MurE in different Bb strains, biochemical evidence show that L-Ornithine is incorporated as the third amino acid in the pentapeptide which is key to the interaction of BpiP with borrelial PG [2,25].
The significance of the presence of L-Ornithine as the third amino acid has implications not only for the interactions of BpiP with the PG but also in the sensitivity to host-and pathogen-derived PG degrading enzymes generating muropeptides capable of stimulating Pathogen Recognition Receptors (PPRs) in the mammalian host. Previous studies have shown that borrelial muropeptides stimulate human Nod2 reporter cell line and a specific inhibitor of the Nod2 signaling pathway inhibited its activation as measured by NF-κB activation [25,62]. It is feasible that PG-binding proteins, like BpiP, while presumably anchoring PG to the inner membrane could limit access to endopeptidases that destabilize the structural integrity of the cell wall. Glycosidases, amidases and peptidases are three major classes of pathogen-encoded proteins that cleave bonds within PG. Compared to 5 amidases-enzymes that cleave polymeric peptidoglycan-present in E. coli (Ami A, B, C, D and AmpD), Lyme spirochetes have been shown to encode for one amidase, N-acetylmuramyl-L-alanine amidase (MurNac LAA) capable of hydrolyzing the amide bond between N-acetylmuramic acid and the N-terminal L-alanine residue of the stem peptide [63]. Moreover, muropeptides accumulated in the medium due to the absence of homologs that contribute to recycling induced Nod2-dependent signaling leading to expression of select cytokines [25,62]. Two putative endopeptidases classified as M23 peptidase domain-containing proteins with similarity to murein DD-endopeptidase (MepM) and murein hydrolase activator NlpD with a LysM domain are present in the genome of Bb B31-A3 strain. However, there are no reports on the role of these proteins in the cleavage of borrelial pentapeptide crosslinks [2,64,65]. Since the in vitro phenotype of bpiP mutant was similar to control strains, it is unlikely that pathogen-derived enzymes impacted bpiP mutant while host-derived PG degrading enzymes could play a role in bpiP survival in mice.
The cell wall biogenesis begins via mevalonate pathway with acetate being converted acetylphosphate and subsequently to acetyl coA by acetate kinase (AckA) and phosphate acetyl transferase (Pta), respectively, [19,22,23]. In addition to driving the mevalonate pathway, levels of acetate (and other Short Chain Fatty Acids (SCFAs) such as butyrate and propionate) induce metabolic changes that alter the levels of several lipoproteins on the surface of the outer membrane via a relatively undefined acid-stress response that drive gene expression by modulating levels of sigma factor, RpoS [23]. The metabolic overlap mediated by host-derived signals/nutrients (such as SCFAs) during initiation of cell wall biogenesis and levels of major surface exposed lipoproteins imply integration of multiple pathways for remodeling the cell envelope/cell wall consistent with the adaptive capabilities of Bb in different hosts [19][20][21]. Transcriptional levels of bpiP were elevated in fed nymphs and under in vitro conditions mimicking fed nymphs (Fig 4) suggesting the possibility of levels of BpiP coinciding with PG synthesis or cell division [36].
The mutant lacking bpiP had no significant differences in in vitro growth rates under conditions mimicking different stages of tick infection (S4 Fig) or in the levels of select borrelial proteins in the mutant (Fig 6). However, the burden of bpiP mutant was significantly lower in immunocompetant C3H/HeN mice at day 14, 28 or 62 post-infection (Fig 7) while viable spirochetes at day 28 and 62 post-infection were not readily detected in infected tissues suggesting a role for the adaptive immune response contributing to the susceptibility of mutant spirochetes (Table 1 and Fig 7B and 7C). While complement-dependent killing of mt spirochetes was also observed with day 14 mouse serum (Fig 15D), most of the tissues from C3H/HeN mice had viable spirochetes at day 14 post-infection although there was a significant reduction in number of mt spirochetes (Fig 7A). The ability of bpiP mt to survive in the immunodeficient SCID mice but not in the immunocompetent BALB/c mice pointed to adaptive immune response of the host in the susceptibility of bpiP mt (Table 2 and Fig 8A and 8B). The likelihood of loss of Lp28-1 encoding for VlsE was also ruled out as wt, mt and ct strains had Lp28-1 and expressed similar levels of VlsE on the surface (Figs 16 and 17). It is interesting to speculate that the lack of BpiP allows for the modulation of the host immune response to be amplified against unknown borrelial determinants that are differentially expressed between mt and the control strains impacting survival of bpiP mutant in immunocompetent mice.
There was a significant reduction in the mt burden in fed infected larvae and fed infected nymphs compared to the levels of wt and ct strains (Fig 9B and 9C). The ability of the naïve larvae to acquire mt spirochetes within 5 days post-infection from needle challenged mice was diminished and only two of 9 infected larvae out of the total 9 examined under the dark-field microscopy were positive for spirochetes and these observations were consistent with the quantitative real-time PCR analysis used to determine the mt burden. These findings reflect the reduced burden of bpiP mt in needle-infected C3H/HeN mice that limited the larvae from acquiring the mt at levels similar to wt or ct strains. These observations, however, were xenodiagnostic of infection in C3H/HeN mice where presence of Bb is detected by the feeding I. scapularis larvae (Fig 9B and 9E) [66,67]. Moreover, there was a significant reduction in the mt burden in fed infected nymphs compared to those infected with wt or ct strains suggesting a lack of expansion of the mt spirochetes following ingestion of a blood meal by infected flat nymphs (Fig 9C). Although, one out of the three mice was positive for infection with the mt in select tissues, collectively the rates of transmission following challenge with infected nymphs was significantly lower for bpiP mt (Table 3 and Fig 9D). These studies demonstrate that absence of bpiP limits the survival of Bb in the mammalian hosts resulting in reduced acquisition/transmission of mt spirochetes via tick larvae and nymphs, respectively. The role of bpiP in the mammalian phase of infection can, therefore, be exploited to reduce the acquisition and transmission of Bb via ticks to naïve mammalian hosts.
We tested the effects of several stressors to determine the underlying basis for the inability of bpiP mutant to survive within mammalian hosts. Vancomycin was previously shown to be an antibiotic that reduces the cell wall stiffness/elasticity impacting shape and motility of Bb [46,47]. Although the timing of BpiP synthesis and interaction with borrelial PG is unclear, bpiP mutant had increased sensitivity to vancomycin at 0.5 and 1μg/ml compared to the control strains (Fig 10A) and lower cfus/ml with 0.5 μg/ml of vancomycin (Fig 10B). Similarly, the bpiP mutant was sensitive to higher concentration of NaCl (200mM) in the BSK growth medium (Fig 11A) with reduced cfus/ml compared to the control strains ( Fig 11B) and exhibit detectable morphological changes such as blebbing compared to the control strains. One possible explanation for this phenotype is the decreased integrity at the site of BpiP-PG interaction within the pentapeptide bridge altering the sensitivity to osmotic stress or morphology of spirochetes in the presence of increased levels of NaCl. Both these stressors did not completely inhibit the bpiP mutant suggesting that the changes in the cell wall stiffness/motility or ability to withstand osmotic stress alone is insufficient to contribute to the reduced infectivity of bpiP mutant.
We hypothesized that mediators of innate immune response due to macrophages, dendritic cells, among other cells, contribute to the significant decrease in the mt burden in C3H/HeN mice at day 14 post-infection before a robust adaptive immune response is established. Lysosomal extracts from bone marrow derived dendritic cells supplemented in borrelial growth medium reduced growth of bpiP mt significantly compared to the parental controls (Fig 12A  and 12B). Moreover, antimicrobial peptide (AMP), LL37, the only member of the human cathelicidin family, significantly reduced the survival of bpiP mt compared to the control strains (Fig 13A and 13B). The mouse ortholog of human LL37, mCRAMP, had no significant effect on the bpiP mutant (S2 Fig) while it was effective against E. coli. These observations were similar to the lack of sensitivity of p66 mutant in the presence of mCRAMP under in vitro growth conditions [68]. Since the positive charge of LL37 favors interaction with negatively charged phospholipids, it is feasible that the loss of bpiP impacts the borrelial cell envelope to facilitate the effects of LL37 on borrelial inner membrane [69]. Both macrophages and dendritic cells exhibited increased uptake of bpiP mutant suggesting that absence of BpiP results in changes that lead to increased uptake (Fig 14B and 14D). Mutants lacking key regulators of gene expression such as BadP or proteins critical for initial stages of colonization of mammalian host exhibit attenuation of infection due to enhanced phagocytosis and killing by macrophages and dendritic cells at the site of inoculation or infection [41]. However, the bpiP mutant had no growth deficit in SCID mice suggesting that soluble and cellular mediators of innate immunity is most likely insufficient to impact colonization while a significant growth deficit could be readily observed in immunocompetent mice as the adaptive immune response gets established. While the levels of Bb-specific antibody responses were similar against wt, mt and ct strains at day 14, it was lower at day 28 and 62 in mice infected with the mt consistent with the lack of survival of bpiP mutant (Fig 15A-15C). It is possible that the complementdependent killing or the inability to withstand the effects of borrelicidal antibodies likely contribute to lack of survival of mt in immunocompetent mice (Fig 15D). It is also feasible that the lack of BpiP reduces the threshold of resistance to multiple antibacterial effectors targeting the PG cell wall and cell envelope of Bb. It has been shown patients with Lyme arthritis develop antibodies specific to borrelial PG [25]. It is likely that the absence of BpiP increases the borrelicidal effects of antibodies either alone or in conjunction with other mediators of innate immunity in mouse models of infection.
Host-derived enzymes that target the borrelial PG during the tick and mammalian phases of infection are being studied in greater detail [15,64]. The immune deficiency (IMD) pathway and Toll pathway in the arthropods respond to diaminopimelic acid [70] and Lysine-type PG present in Gram-negative and Gram-positive bacteria, respectively [15,71,72]. Unlike DAPtype PG, it is unclear how the L-Ornithine type PG in Bb is sampled to activate the IMD pathway [15]. Similarly, there are 4 homologs of Peptidoglycan Recognition Proteins (PGRPs) in mice and humans with variable levels of expression, predicted domains and cellular locations that could contribute to the host response to borrelial PG [73][74][75]. Moreover, a lipoprotein, LipL21, bound and protected leptospiral PG from being digested as muropeptides prevented recognition of Leptospires via Nod1 or Nod2 signaling pathways although BpiP does not appear to be a lipoprotein [76].
In this study, we demonstrate the importance of BpiP in the contribution of cell wall/cell envelope in the survival of Bb within the mammalian hosts [77]. As summarized in Fig 16, the organization of the borrelial cell envelope and the unique composition of borrelial PG confer adaptive capabilities for Bb during the infectious cycle. Our studies show that BpiP plays a key role in the ability of Bb to survive in the mammalian host. Moreover, the nature of innate and adaptive immune response mounted against surface-exposed borrelial determinants and PG breakdown products impact the clearance or persistence of spirochetes in mammalian hosts (Fig 18). The molecular mechanisms that confer integrity to borrelial cell wall/cell envelope mediated via proteins like BpiP are likely to sustain the survival of Lyme spirochetes during difference phases of the infectious cycle. Strategies that subvert these mechanisms are likely to aid in developing novel preventive and therapeutic options for control of Lyme disease.

Ethics statement
All animal experiments were conducted following NIH guidelines for housing and care of laboratory animals and in accordance with protocols approved by the Institutional Animal Care and Use Committee (protocol number MU071) of the University of Texas at San Antonio (UTSA). Based on these guidelines, general condition and behavior of the animals were monitored by trained staff. The animal facilities at UTSA are part of Laboratory Animal Resources Center (LARC), which is an AAALAC International Accredited Unit. During tick infection phase, all animals were monitored every 12 hours and tick-challenged animals were housed under Biosafety Level 2 facility with restricted entry.

Bacterial strains and growth conditions
A low passage infectious clonal isolate of Borrelia burgdorferi B31-A3 was used as the parental strain in all the studies (Table 4) [78]. Wild type (wt), mutant (mt) and complemented (ct) strains were propagated in liquid Barbour-Stoenner-Kelly (BSK II) media supplemented with 6% heat inactivated rabbit serum (Pel-Freez Biologicals, Rogers, AR) with appropriate antibiotics (Sigma-Aldrich, St. Louis, MO) as previously described [12,19,20,[38][39][40]. All borrelial strains were first grown at pH7.6/32˚C (laboratory growth conditions) and when the cultures reached a density of 1×10 7 spirochetes/ml, the bacteria were seeded into fresh BSK II medium at pH 7.6/23˚C to mimic the midgut of un-fed tick. Once the cultures at un-fed tick mimicking conditions reached a density of 1-2×10 7 spirochetes/ml, the spirochetes are seeded into fresh BSK II medium at pH 6.8/37˚C mimicking the midgut of fed ticks. Total RNA or protein lysates were extracted/prepared from these cultures when the density of Bb reached 5x10 7 cells or 1x10 8 Bb /ml for transcriptional and translational analyses, respectively. Harvested cells were washed three times with Hanks Balanced Salt Solution (HBSS, HyClone, ThermoFisher Scientific, Waltham, MAA) and processed for extraction of RNA or proteins.

Generation of recombinant proteins and antibodies
Total genomic DNA was isolated from a low passage clonal isolate of Bb by phenol chloroform extraction followed by ethanol precipitation as previously described [12,19,[38][39][40]80]. Recombinant proteins and monospecific sera were generated as previously described [12,19,38,40,79,80,82,83]. Briefly, total genomic DNA obtained from B. burgdorferi B31-A3 was used as a template to amplify genes of interest by PCR using primers listed in Table 5. Amplified genes were cloned into the cloning vector pCR2.1 TOPO using the TOPO 2.1 cloning kit (Invitrogen) and transformed into electrocompetent TOP 10 E. coli (Invitrogen, Carlsbad, CA). Positive colonies containing the amplified gene in pCR2.1 were selected via blue/white screening using plates supplemented with IPTG/X-gal followed by restriction enzyme digestion (EcoRI) of the plasmids from positive clones to confirm the presence of the insert of the correct size (New England Biolabs). The resulting plasmids were then digested with BamHI and SalI and cloned into the inducible expression vector pMAL-p2X (New England Biolabs). Site-specific substitutions replacing (1) E315 or (2) with 7 site directed mutations (7SDMs) replacing amino acids of PG-interacting motif of BpiP with alanines or glycine were performed using specific primers as listed in Table 5. The cloned expression vector was transformed into the E. coli expression host, Rosetta (Novagen). Protein expression was induced with 1mM IPTG for 4 hours at 37˚C in a 500ml flask and the culture was pelleted by centrifugation for 10 min at 3000 x g (SORVALL, F13S rotor) followed by resuspension in Low Salt Buffer (LSB, 10mM phosphate buffer, pH7.0 containing 30mM NaCl and 10mM β-mercaptoethanol). Resuspended cells were then lysed using a French press. The resulting lysate was then spun down at 4,000 x g for 30 minutes and supernatant was loaded on to a column with amylose beads (New England Biolabs) and purified using the NGC chromatography system [84]. The column was washed with buffer A (LSB) and proteins were eluted with buffer B (LSB containing 50mM maltose). Fractions were analyzed on a 12% SDS PAGE gel. Maltose binding protein (MBP) was cleaved from purified recombinant BpiP proteins (wild type, E315A or 7SDMs) in LSB by incubating with TEV protease tagged with 7X Histidine at 4˚C, overnight in TEV protease buffer (50mM Tris-HCl, 0.5mM EDTA, 1mM DTT at pH 7.5). TEV protease was removed using NEBExpress Ni Spin column as per the manufacturer's protocol and MBP was then removed by repeated passage over amylose beads (New England Biolabs). The top aqueous fractions with only the BpiP proteins were collected, concentrated using centrifugal filter devices (Millipore) and separated on 12% SDS-PAGE gel, transferred to PVDF membrane. Immunoblot analysis was performed to evaluate the efficacy of cleavage of the MBP using antibodies directed against either MBP or BpiP (S1 Fig).
Recombinant C-terminal region A260-N380 of BpiP cloned into plasmid pMCSG7 with an N-terminal His-tag (BbOmpA CTD ) was expressed and purified using Ni-NTA column [30] and the His-tag cleaved with TEV proteas. Recombinant C-terminal region of BpiP without His-tag was then emulsified in equal volume of Titermax gold adjuvant, respectively and used to immunize six to eight weeks old BALB/c mice (n = 3) with 50 micrograms/mouse (day 0). Boosters were administered at day 14, and 21 and serum was collected on day 28.

Isolation of Peptidoglycan (PG) from B. burgdorferi
Wild type Bb (B31-A3) was grown in BSKII medium (1L) supplemented with 6% normal rabbit serum at pH 7.6/32˚C to a density of 1x10 8 Bb/ml and harvested by centrifugation. The pellet was resuspended in 20 ml ice-cold ultrapure water and was added to an equal volume of 8% SDS solution stirred in a beaker placed within a boiling water bath with continuous stirring for 1 hour [85][86][87][88]. After cooling to room temperature, PG was pelleted by ultracentrifugation at 130,000 x g using SORVALL Discovery 90 with Surespin 630 rotor at room temperature for 1 hour. The PG pellet was then resuspended in 20 ml of ultrapure water with vigorous vortexing and added to boiling 8% SDS solution and incubated for 15 min followed by ultracentrifugation at 130,000 × g at room temperature for 1 hour. After washing the PG with ultrapure water four times and pelleting by ultracentrifugation, PG was treated with 0.1 mg/ml α-amylase at 37˚C for 2 hours followed by treatment with 0.2mg/ml pronase for 90 minutes at 60˚C to hydrolyze proteins associated with PG. The mixture was added to an equal volume of boiling 8% SDS solution and incubated for 15 min in boiling water bath. The final pellet was then washed four times with water as described above and pelleted at 130,000 × g at room temperature for 1 hour [89]. The final PG pellet was stored at 4˚C for further analysis.

Muramic Acid Assay for Determining PG concentration
Purified peptidoglycan (40 μl) was resuspended in ultrapure water (Sigma) and an equal volume of 5M H 2 SO 4 solution was added and incubated at 90˚C for 2 hours to hydrolyze the PG as described previously [89]. Muramic acid standard solutions (0-1 mM) were treated similarly to generate a standard curve. After incubation, 180μl of ultrapure water and 90μl of 10M NaOH solution were added and incubated at 37˚C for 30 mins to release lactic acid from muramic acid residues of PG. Then transfer 80 μl of hydrolyzed samples to clean glass tubes in triplicates and add 500 μl concentrated H 2 SO 4 (18.8 M). Incubate glass tubes in a boilingwater bath for 5 mins and after tubes cooled down to room temperature, five μl of 4% CuSO4 solution and 10 μl 1.5% 4-phenylphenol solution were added and samples were incubated at 30˚C for 30 mins and absorbance measured at 570nm. Peptidoglycan muramic acid units were determined using standard curve.  [85,86,89]. An equal volume of Buffer B (50mM sodium phosphate [pH8.5] containing DTSSP (10mM) on ice for 30 mins was added followed by centrifugation at 21, 000 x g to pull down the cross-linked complexes. The samples were mixed with 2X SDS-PAGE sample buffer with 10% β-mercaptoethanol (β-ME) and boiled at 95˚C for 5 minutes and the supernatant were analyzed on SDS-12% PAGE gel and transferred to PVDF membranes. The membranes were developed with mouse antiserum directed at C-terminal region of BpiP, goat anti-mouse IgG conjugated to HRPO and developed using ECL. BpiP-PG crosslinked complexes are unable to enter the gel but the cleavage of the cross-linker by β-mercaptoethanol treatment prior to electrophoresis facilitates analysis of the interactions of wild type BpiP or BpiP with site-specific changes with borrelial or heterologous PG in the supernatant as described in the flow chart in S3 Fig [86,87,89].

Partitioning Bb cell envelope
The isolation of Bb outer membrane vesicle (OMV) and protoplasmic cylinder (PC) were isolated using a modified method as described previously [90]. Briefly, 1x10 9 Bb cell were harvested by centrifugation and washed twice by 1x PBS (pH 7.4) and the pellet was re-suspended in 30 ml of ice-cold 25mM citrate buffer (pH 3.2). The mixture was incubated at room temperature for 2 hours and vortexed for 1 min every 30 mins. After incubation, the cells were pelleted at 20,000 x g for 30 mins and re-suspended in 3 ml of 25mM citrate buffer and layered onto a discontinuous sucrose gradient which composed of 56%, 42% and 25% sucrose. Then the whole gradient was centrifuged at 100,000 x g for 16 hours at 4˚C using Beckman SW28 swing bucket. After centrifugation, the OMV fraction (upper band) and the PC fraction (lower band) were removed and the final fractions were stored in SDS loading buffer and used for further immunoblot analysis. Periplasmic and spheroplasmic fractions of Bb expressing pYC153 as reported previously [91,92]. Briefly, 2x10 8 Bb were treated with periplasting buffer (200mM Tris-HCl pH7.5; 50mM KCl, 1mM EDTA and 30units/μl of Ready-Lyse Lysosyme (Lucigen) and Omnicleave endonuclease. The cells were treated with sterile water and centrifuged to obtain the periplasmic fraction and the pellet treated with Peripreps Lysis buffer (10mNM Tris-HCl (pH7.5), 50mM KCl, 1mM EDTA and 0.1% dexoycholate, 10mM MgCl 2 ). The supernatants were collected as spheroplasts. All samples were analyzed by immunoblot analysis using mouse anti-FLAG-tag, rabbit anti-Green Fluorescent Protein (gfp) and anti-FlaB monoclonal antibodies.

Generation of bpiP mutant, cis-and trans-complemented strains
The low passage Bb B31-A3 strain described above was used for all genetic manipulations. Deletion of bpiP was performed as previously described [38][39][40]79]. A suicide vector that contained homologous regions spanning 1,000 base pairs up-and down-stream of bpiP was used to replace the native copy of bpiP gene by allelic exchange with selectable streptomycin resistance marker under the control of the constitutive borrelial promoter P flgB . The up-and downstream regions were amplified using primers listed in Table 5 and cloned into pCR2.1 cloning vector. The two fragments were then cloned together using the engineered restriction enzyme sites on the primers and those present in the pCR2.1 vector. An engineered Sal1 site between the two fragments was used to clone the streptomycin resistance marker as described above.
The plasmid DNA was linearized with Sca1 to facilitate homologous recombination via double cross over events following electrotransformation of Bb B31-A3 strain as described previously [43,78,89,93,94]. Mutant colonies of Bb were obtained by plating in BSKII agar overlays in the presence of appropriate concentrations of antibiotics and colonies were screened by PCR and plasmid profile determined to be identical to parental strain. Complementation of bpiP was performed using the same methods described above except the primers were designed for both upstream region and the bpiP gene. To select for proper complementation, a gentamicin resistance marker, aacC1 under the control of the P flgB promotor, was cloned downstream of the bpiP gene.
The trans-complemented strain was generated by transforming the bpiP mutant using a borrelial shuttle vector pYC153 obtained by cloning 1kb upstream of bpiP and bpiP-3XFLAGtag at the Kpn1 site present in the plasmid pTM61 [44].
Colonies were isolated and screened in the same manner as described above for bpiP mutants. All Bb strains generated as part of this study subjected to whole genome sequencing at the UTSA genome sequencing core and were found to have identical plasmid profile. Analysis of the sequence of the borrelial chromosome confirmed the location of the deletion/cis-complementation consistent with the region within plasmid constructs used to generate these strains.

Sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) and Immunoblot analysis
Total borrelial lysates were obtained by harvesting the spirochetes as described above followed by separation on an 12% SDS PAGE gel. To ensure equal loading the gel was either stained with Coomassie Brilliant Blue (ThermoFisher) or transferred to PVDF membrane and probed with antibodies specific for borrelial FlaB as previously described [45,82]. Proteinase K treatment of spirochetes was done as reported previously [95] Immunoblot analysis was performed as previously described with monospecific antisera from mice, rats, or rabbits generated against various borrelial proteins [41]. Blots were developed using appropriate HRP-conjugated secondary antibody (GE Healthcare) and ECL substrate (Thermo Fisher) followed by exposure to X-ray film.

in vitro growth phenotype of borrelial mutants
Wt, mt and ct strains were diluted from cultures at pH7.6/32˚C and re-seeded at 5×10 5 spirochete/ml in fresh BSK II medium under growth conditions reflecting laboratory (pH 7.6/ 32˚C), tick mid-gut before (pH 7.6/23˚C) and after (pH 6.8/37˚C) a blood meal. Cultures were enumerated every 24 hours using dark field microscopy using motility as a measure of viability. The cultures were grown in triplicate, with three independent biological samples. Error bars indicate standard error. Levels of significance were determined using unpaired t test with GraphPad Prism software (Graph-Pad Software, Inc., San Diego, CA) and a p value of less than 0.05 was considered significant.

Complement-dependent killing of Bb
The anti-borrelial serum used in this study was from mice infected with the mt strain (n = 9; 14 days post-infection), and the borrelicidal activity was determined as described previously [96][97][98]. Briefly, 50 μL of diluted mouse serum (1:80) was mixed with 10 μL of guinea pig complement (guinea pig complement, Sigma-Aldrich, # S1639) or heat-inactivated guinea pig complement (negative control) with wt, mt or ct strains (5×10 6 spirochete/ml) in 40 μL of BSK II (pH 7.6) medium and then incubated at 32˚C for 24 h. Motile spirochetes following treatment were enumerated using dark-field microscopy. The survival percentage was calculated using the numbers of motile spirochetes from serum-treated samples to those with heat-inactivated complement (negative control). Error bars indicate standard error. Levels of significance were determined using unpaired t test with GraphPad Prism software and a p value of less than 0.05 was considered significant.

Sensitivity assays
Antibiotics. Wt, mt and ct strains were diluted from stationary phase (1×10 8 spirochete/ ml) cultures (pH 7.6/32˚C) and then re-seeded at 1×10 6 spirochete/ml in fresh BSK II medium supplemented with vancomycin at pH 7.6/32˚C in 96 well plate. Motile spirochetes were enumerated after 10 days of incubation. Colony forming units of cultures with 0 or 0.5 μg/ml of vancomycin at day 5 post-treatment were determined by diluting cultures of all three strains. The liquid cultures of all three strains were diluted 1:100,000 in 1X BSKII medium and plated as agar overlays with BSKII media with 0.84% agarose and the final CFU/ml were calculated after 10 days of incubation at 32˚C with 1% CO 2 .
Osmolarity/ salt concentration. Wt, mt and ct strains were diluted from stationary phase (1×10 8 spirochete/ml) cultures (pH 7.6/32˚C) and then re-seeded at 1×10 6 spirochete/ml in fresh BSK II medium supplemented with different concentrations of NaCl at pH 7.6/32˚C. Cultures were enumerated every 24 hours using dark field microscopy. The liquid cultures of all three strains were diluted 1:100,000 in 1X BSKII medium and plated as agar overlays in BSKII media with 0.84% agarose and the final CFU/ml were calculated after 10 days of incubation at 32˚C with 1% CO 2 . The cultures were grown in triplicate, with three independent trials. Error bars indicate standard error. Levels of significance were determined using unpaired t test with GraphPad Prism software and a p value of less than 0.05 was considered significant.
Mouse and tick infections. Unless stated otherwise, mice were infected by needle inoculation via intradermal route as previously described [40]. Six to eight weeks old C3H/HeN, BALC/c or SCID mice were injected intradermally with 10 5 spirochetes per mouse. At 14, 28 or 62 days, post-infection, mice were euthanized and a piece of the abdominal skin, spleen, bladder, heart, left inguinal lymph node, and left tibiotarsal joint was crushed and cultivated in BSK-II media to isolate Bb. Cultures were scored for the presence of spirochetes 2 to 3 weeks later using dark field microscopy after one blind passage at day 5 post-isolation. Another piece of abdominal skin, half of the spleen, right inguinal lymph node and right tibiotarsal joint were used to isolate total DNA using High pure PCR template preparation kit (Roche) for qPCR analysis to enumerate the number of spirochetes in each tissue. Infectivity analysis using BALB/c and SCID mice were done at day 28 post-infection to determine the effects of adaptive immune response on the colonization capabilities of the bpiP mt.
Ixodes scapularis larvae were purchased from the tick rearing facility at Oklahoma State University (Stillwater, OK) and incubated in a chamber at 22˚C with 99% humidity and a 15-hour light / 9-hour dark cycle. Larvae were allowed to feed on infected C3H/HeN mice with wt, mt or ct strains. Nine fed larvae were crushed to assess spirochete burden. The others were allowed to molt into nymphs and nymphs were then allowed to feed on naïve C3H/HeN mice to repletion (3-5 days). Fed nymphs recovered after feeding were washed with 3% H 2 O 2 , 70% ethanol and ddH 2 O sequentially and crushed with a small volume of BSKII medium and inoculated into 3ml of fresh BSK-II medium or used for extracting total genomic DNA. The media containing crushed ticks was then scored 1 to 2 weeks later for the presence of spirochetes by dark field microscopy. Recovered spirochetes were subjected to PCR analysis for confirmation of correct strains recovered from ticks. Total genomic DNA was extracted from fed ticks followed by quantitative real time PCR using primers specific for Bb flaB. The flab copies per tick was graphed and the difference between Bb colonization was determined using unpaired t test with GraphPad Prism software. A p value of less than 0.05 was considered significant.

Killing assay with crude lysosomal extracts
C3H/HeN or C57BL/6 mice were used in these experiments and housed under pathogen-free conditions under institutionally recommended guidelines of Laboratory Animal Resource Center at The University of Texas at San Antonio. Murine bone marrow-derived dendritic cells (BMDC) cultures were established as described previously [99]. Briefly, bone marrow cells were flushed from the femurs and tibiae of C57BL/6 mice. BMDC were cultured in RPMI 1640 (Gibco Life Technologies) supplemented with 10% heat-inactivated fetal bovine serum (iFBS), 2 mM L-glutamine, 100U/ml penicillin, 100μg/ml of streptomycin, and 50 mM β-mercaptoethanol (complete medium). Cells were washed, counted, and plated in complete medium supplemented with 20ng/ml GM-CSF (PeproTech). One half of the medium was replaced every three days, and the cells were harvested on day 8. The cells were then purified by negative selection of F4/80 + cells using F4/80 biotin and anti-biotin beads (Miltenyi Biotec, Auburn, CA). The flow-through cells were further positively selected using magnetically labeled CD11c antibodies (Miltenyi Biotec).
Lysosomal extracts were prepared as described previously [99]. Lysosomal extraction buffer (1X Sigma-Aldrich) was added at 2ml per 3×10 8 cells. DCs were homogenized with a Power-Gen 700 homogenizer (Fisher scientific, Pittsbrugh, PA) using a 7-by110-mm homogenizer tip (Fisher Scientific). The homogenizer was passed through the cells 20 to 25 times. Cells were centrifuged at 1000 x g 10 min to remove intact cells and cellular debris. Then the supernatants were collected and centrifuged for 20 min at 20,000 x g to pellet lysosomes. The pellet containing lysosomes was resuspended in 1 mL of extraction buffer. The sample was then sonicated for 20 s at a setting of 40% on a model 500 Sonicator (Fisher scientific) to obtain the final crude lysosomal extract.
Wt, mt and ct spirochetes grown in BSKII media were harvested (1x10 6 Bb) and resuspended in 1mL of new BSKII media (pH 6.8) supplemented with or without 10% lysosomal extract and incubated at 37˚C, 1% CO 2 for five days. The number of viable spirochetes was determined every 24 hours using dark field microscopy. The liquid cultures of all three strains were diluted 1:100,000 in 1X BSKII medium and plated as agar overlays in BSKII media with 0.84% agarose and the final CFU/ml were calculated after 10 days of incubation at 32˚C with 1% CO 2 . Levels of significance were determined using unpaired t test with GraphPad Prism software and a p value of less than 0.05 was considered significant.

Imaging flow cytometry to measure uptake of Bb by macrophage and dendritic cells
Uptake of wt, mt and ct strains of bpiP by macrophages and dendritic cells were performed using the ImageStream1X Mark II Imaging flow cytometer. Activated macrophages were obtained from six to eight-week old C3H/HeN mice after intraperitoneal injection of 1 ml of sterile 3% proteose peptone as previously described [79]. The mice were euthanized by CO 2 asphyxiation after 4 days and 20 ml of DMEM (GIBCO) was injected to flush the cells from the peritoneal cavity. The cells were spun down and treated with 0.14M ammonium chloride to remove red blood cells followed by resuspension in DMEM supplemented with 10% iFBS. Cells were enumerated and then incubated with CFSE labeled Bb strains at a ratio of 1:1 in 50% cell culture medium (DMEM +10%iFBS) and 50% borrelial growth medium (BSK-II with 6% inactivated normal rabbit serum) at 37˚C in 5% CO 2. Cells were then spun down and resuspended in 2% paraformaldehyde (PFA) and sequentially washed two times with FACS buffer (PBS with 5% FBS). After washing, cells were resuspended in FC block at a 1:250 dilution (ThermoFisher) for 15 minutes. After incubation, the cells were washed twice with FACS buffer, resuspended in 50μl of FACS buffer and labeled with 0.8 μl of anti-CD45 PE, 0.4 μl of anti-F4/80 APC, 1.5 μl of anti-CD11C BV421 and 1 μl CD11b PE-Cy7 (ThermoFisher) and incubated at 4˚C in the dark for 30 minutes. After incubation, the cells were washed twice with FACS buffer, resuspended in 2% PFA and samples were evaluated using the ImageStream1X Mark II Imaging flow cytometer at the UTSA Immune Defense core [45].
In order to enumerate the number of macrophages in the total cell population that were positive for Bb uptake, the total single cell population was first gated for CD45 + F4/80 + cells. This subset population was then gated for the percentage of cells that were also CFSE positive indicating cell with Bb. In order to determine the number of dendritic cells in the total cell population that were positive for Bb uptake, the total single cell population was first gated for CD45 + F4/80cell, then the subset population was then gated for CD11C + cells. The remaining population was gated for the percentage of cells that were also CFSE positive due to Bb infection. Internalization of the spirochetes in macrophages and dendritic cells were determined using the Aminis IDEAS software suite.

DNA sequencing for NGS
Sequencing of genomes was performed as previously described [41,100]. Briefly, total genomic DNA from each strain was obtained through phenol chloroform extraction and checked for purity spectrophotometrically using NanoDrop microvolume spectrophotometer (Thermo-Fisher). Total gDNA was then submitted to the UTSA genome sequencing core for library preparation and sequencing on an Illumnia MiSeq. Resulting files were then uploaded to basespace and contigs were assembled using the velvet assembly program [101]. Contigs were then mapped to the reference chromosome and plasmids from NCBI using Geneious [2].

S1 Fig. Purification of recombinant BpiP proteins.
Recombinant BpiP proteins (wild-type and with site-specific changes) fused to maltose binding protein (MBP) with an intervening TEV protease site were overexpressed and purified using amylose column in conjunction with FPLC (BIORAD). MBP was cleaved from BpiP using recombinant TEV protease fused to 7X-His-Tag bound to NiNTA beads. Recombiant BpiP was then concentrated from the flowthrough fractions and was separated on 12% SDS-PAGE gel, transferred to PVDF membranes. Immunoblot analysis was performed using anti-Maltose serum or anti-BpiP serum generated against C-terminal region of BpiP followed by goat-anti-mouse IgG conjugated with HRPO and blots developed using Enhanced Chemiluminescence. Lane 1) BpiP-MBP, 2) BpiP-MBP + TEV treatment, 3) BpiP only, 4) E315A BpiP-MBP, 5) E315A BpiP-MBP +TEV treatment, 6) E315A BpiP only, 7) 7SDMs BpiP-MBP, 8) 7SDMs BpiP-MBP + TEV treatment, 9) 7SDMs BpiP only. (TIF) S2 Fig. Sensitivity of bpiP mutant to murine antimicrobial peptides. All three strains (wt, mt and ct) were propagated at 10 5 /ml in BSKII growth medium at pH 6.8/32˚C with 100 μg/ ml of mCRAMP. Cells were enumerated every 24 hours using dark field microscopy. The cultures were grown in triplicate and one out of two independent experiments is shown. Statistical analysis of the difference in the number of wt and mt or ct and mt spirochetes was done by unpaired t test. Wild type (B31/A3), mt and ct strains were diluted from stationary phase (1×10 8 bacteria ml −1 ) cultures, re-seeded at 5×10 5 bacteria ml −1 in BSKII medium and enumerated every 24 hours using dark field microscopy under different growth conditions. The cultures were grown in triplicate, with three independent trials. Error bars indicate standard error. Levels of significance were determined using two-way ANOVA with α = 95%. (TIF)