Control subjects who were asymptomatic without a history of Lyme disease and patients with a history of Lyme disease were recruited for the study after written informed consent to collect and publish their data was obtained. Approval for sample collection was obtained from the Western Institutional Review Board, Olympia, WA (WIRB ^® #20141439). Further approval for sample testing was obtained from the Institutional Review Board of the University of New Haven, West Haven, CT. Serological testing of all participants after coding of their blood samples was performed by IGeneX Reference Laboratories, Palo Alto, CA in a blinded fashion.

Patients were considered positive for Lyme disease if they were serologically positive by CDC criteria and/or IGeneX criteria, as previously described ( Engstrom et al., 1995; Ma et al., 1992), or if they had musculoskeletal, neurocognitive and/or cardiac symptoms clinically consistent with a Lyme disease diagnosis, as described elsewhere ( Donta, 2014; Smith et al., 2014). None of the patients were taking antibiotics at the time of testing.

Borrelia spirochetes were cultured as previously described ( Bankhead & Chaconas, 2007; Middelveen et al., 2013b; Middelveen et al., 2014a). The inoculum for blood culture was prepared as follows: 10 milliliters of whole blood was collected by sterile venipuncture from each patient. Samples sat at room temperature for 10 to 15 minutes allowing clotting to occur. Red blood cells (RBCs) were separated by low speed centrifugation. Barbour–Stoner–Kelly H (BSK-H) complete medium was used for cultures with the addition of 6% rabbit serum (Sigma Aldrich, #B8291) and the following antibiotics: phosphomycin (0.02 mg/ml), rifampicin (0.05 mg/ml), and amphotericin B (2.5 µg/ml) (Sigma Aldrich).

The culture medium described above was inoculated for blood culture with the spun serum containing white blood cells and some RBCs, and for genital culture with either ejaculated semen or vaginal secretions collected by intravaginal swabbing with a sterile cotton-tipped swab. Blood and genital cultures were incubated at 32°C in an Oxoid anaerobic jar (Thermo Scientific) containing an AnaeroGen sachet (Thermo Scientific) to provide an anaerobic environment. Cultures were incubated for four weeks and checked weekly by light and/or darkfield microscopy for visible motile spirochetes.

All cultures were processed for microscopic imaging and PCR by centrifuging the culture fluid at 15,000 g for 20 minutes to concentrate spirochetes. The supernatant was discarded and the pellet retained. The pellet samples were coded and processed in a blinded fashion for subsequent experiments.

Dieterle silver staining was performed using two fixation methods. In the standard method, formalin-fixed, paraffin-embedded pellets were sectioned and stained with Dieterle silver stain as previously described ( Aberer & Duray, 1991; Middelveen et al., 2013a). In the newer method, culture fluid was spread and dried on a SuperFrost™ Plus microscope slide (Fisher Scientific) and fixed by incubating the slide in acetone for 10 minutes at -20°C, as previously described ( Sapi et al., 2013). Dieterle silver staining was performed on the acetone-fixed slide.

A. McClain Laboratories. Blood and genital culture pellets from coded patient samples were processed in a blinded fashion for special staining at McClain Laboratories LLC, Smithtown, NY. Formalin-fixed, paraffin-embedded pellets were sectioned and stained with Dieterle silver stain or anti-Bb immunostains for spirochete detection, as previously described ( Middelveen et al., 2013a; Middelveen et al., 2014a). In brief, immunostaining was performed using an unconjugated rabbit anti-Bb polyclonal antibody (Abcam ab20950), incubated with an alkaline phosphatase probe (Biocare Medical #UP536L), followed by a chromogen substrate (Biocare Medical #FR805CHC), and counterstained with hematoxylin. Positive and negative controls were prepared for comparison purposes with liver sections from Bb-inoculated mice and uninfected mice followed by Dieterle and immunostaining. Culture pellets from mixed Gram-positive bacteria and mixed Gram-negative bacteria were also prepared for comparison purposes as negative controls to exclude cross-reactivity with commonly encountered microorganisms. Staining was titrated to determine optimal antibody dilutions to achieve positive staining of spirochetes while minimizing background staining ( Middelveen et al., 2013a; Middelveen et al., 2014a).

B. University of New Haven. Coded samples were processed in a blinded fashion for Bb immunostaining as previously described ( Sapi et al., 2013). Culture fluid was spread and dried on a SuperFrost™ Plus microscope slide (Fisher Scientific) and fixed by incubating the slide in acetone for 10 minutes at -20°C. Dried, fixed culture fluid was submerged under 100 μl of polyclonal FITC-labeled rabbit anti-Bb antibody (Thermo Scientific #PA-1-73005) diluted 1:50 in 1× PBS buffer with 1% BSA (Sigma Aldrich #A9418). For negative controls, the antibody was omitted and replaced with normal rabbit serum. The slides were then incubated for 1 hour at 37°C in a humidified chamber, washed with 1× PBS for 5 minutes at room temperature, rinsed twice in double distilled water and dried in a laminar air-flow hood for 10 minutes. The slides were mounted with Vectashield mounting medium with DAPI counterstain (Vector Labs) and viewed with fluorescent microscopy at 400× magnification with a Leica DM2500 microscope ( Sapi et al., 2013).

The Bb molecular beacon DNA probe was generously provided by Dr. Alan MacDonald. Probe FlaB (sequence of 23 mer) was derived from the Bb open reading frame (ORF) BB0147 (approximately 1100 mer) of the flagellin B gene. A nucleotide Basic Local Alignment Search Tool (BLAST) search of the 23 mer sequence disclosed no matches in the human genome or in any other life form other than the Bb sequence of BB0147.

Bb detection with the molecular beacon was performed as previously described ( Middelveen et al., 2014a) on coded samples in a blinded fashion using the following protocol: paraffin sections were dewaxed by baking at 60°C, then immersed in serial 100% xylene baths followed by serial immersion through baths of 100% ethanol, 90% ethanol, 80% ethanol, and finally in distilled H ₂O, and then air-dried. Fixed sections were immersed in 20 μl of the working DNA beacon solution. The sectioned specimen was covered with a layer of plastic cut from a Ziploc ^® freezer bag and was heated at 90°C for 10 minutes to denature DNA and RNA. The heat was first reduced to 80°C for 10 minutes, then the slides were removed from heat and allowed to gradually cool to 24°C. The slides were washed in PBS, covered with 30% glycerol and a glass coverslip, then examined under an EPI Fluor microscope. Staining of test specimens was performed alongside staining of positive and negative controls. The positive control was prepared by embedding a known Bb strain in agarose, formalin-fixing the specimen then blocking in paraffin and staining sections as described above.

The specificity of the FlaB probe was validated in studies performed at the University of New Haven (Sapi E., unpublished observation 2014; see Supplemental Figure 1). The FlaB probe hybridized to Bb sensu stricto, yet failed to hybridize with B. afzelii, B. garinii, B. hermsii, Treponema denticola and Escherichia coli. Thus the probe appears to be specific for detection of Bb sensu stricto.

Blood and genital culture pellets were first dissolved in 200 μl of Qiagen buffer, then forwarded to the University of New Haven, Department of Biology and Environmental Science, West Haven, CT, USA and Australian Biologics, Sydney, NSW, Australia for PCR detection of Borrelia. All control and patient samples were coded, and PCR testing was performed in a blinded fashion.

A. Australian Biologics. Detection of Borrelia by PCR was performed as previously described ( Mayne et al., 2012) using the Eco™ Real-Time PCR system with primers targeted to the genes encoding 16S rRNA ( Borrelia), flA ( T. denticola) and fliG1 ( T. pallidum) and analyzed with the software version 3.0.16.0. DNA was extracted from the dissolved culture pellets using the QIAamp DNA Mini Kit (Qiagen) and 20 μl were used for each reaction. The thermal profile involved incubation for 2 minutes at 50°C, polymerase activation for 10 minutes at 95°C then PCR cycling for 40 cycles of 10 seconds at 95°C dropping to 60°C sustained for 45 seconds. All samples were run in duplicate with positive and negative controls. Positive controls were genomic DNA samples from B. burgdorferi, B. garinii, and B. afzelii (Amplirun DNA/RNA amplification controls, Vircell S.L, Granada, Spain). Negative controls were samples of non-template DNA in molecular-grade water. The magnitude of the PCR signal generated (∆R) for each sample was interpreted as positive or negative compared to positive and negative controls.

In samples with sufficient DNA for sequencing, endpoint PCR amplification and Sanger sequencing of the Borrelia gene target from cultures was followed by BLAST comparison with known Borrelia sequences, as previously described ( Mayne et al., 2012).

B. University of New Haven. DNA samples were extracted from blood, vaginal or seminal cultures by lysing cells overnight in 180 µl tissue lysis buffer (Qiagen) and 20 µl Proteinase K (Qiagen) at 56°C in a shaking water bath followed by phenol:chloroform extraction the next day. The DNA was resuspended in 50–100 µl 1×TE buffer.

A published TaqMan assay targeting a 139-bp fragment of the gene encoding the Borrelia 16S rRNA was used for the detection of Borrelia in DNA extracted from patient samples ( O’Rourke et al., 2013). All reactions were carried out at a final volume of 20 µl and consisted of 900 nM of each primer, 200 nM of probe, and 10 µl of 2× TaqMan Universal PCR Master Mix (Applied Biosystems) and 1 nanogram of DNA. Amplifications were carried out on a CFX96 Real-Time System (Bio-Rad), and cycling conditions consisted of 50°C for 2 minutes, 95°C for 10 minutes, followed by 40 cycles of 95°C for 15 seconds and 60°C for 60 seconds. Fluorescent signals were recorded with CFX96 Real-Time software and Cq threshold was set automatically. The reactions were performed in triplicate with positive and negative controls.

Nested PCR primers for the genes encoding the Borrelia 16S rRNA, fla and pyrG loci were used as previously described ( Clark et al., 2013; Margos et al., 2010; Sapi et al., 2013). Reactions were carried out in a final volume of 50 µl using 10 µl template DNA. Final concentrations were 2× Buffer B (Promega), 2 mM MgCl ₂, 0.4 mM dNTP mix, 2 µM of each primer, and 2.5 U Taq polymerase (Invitrogen). “Outer” primers were used in the first reaction. “Inner” primers were used for the nested reaction, in which 1 µl of PCR product from the first reaction was used as template for the second. Cycling parameters were as follows: 94°C for 5 minutes followed by 40 cycles of denaturation at 94°C for 1 minute, annealing for 1 minute (temperature based on the primer set used), and extension at 72°C for 1 minute, with a final extension step at 72°C for 5 minutes. PCR products were visualized on 1–2% agarose gels. Sanger sequencing was used for gene analysis, as previously described ( Margos et al., 2010).

All patient data are shown in Table 1. The control group included four asymptomatic patients (two males and two females). All four were seronegative for Bb.

The patient group included six male subjects and seven female subjects, including four pairs of partners (Patients 6 and 7, 8 and 9, 10 and 11, and 12 and 13, respectively). Eleven of the 13 patients selected for the study were serologically positive for Lyme disease. Patient 1 was serologically equivocal and patient 8 was seronegative, although Bb plasmid DNA was detected in whole blood and serum from this patient.

Blood cultures from 11 patients were incubated for four weeks and checked weekly for spirochete growth using light and darkfield microscopy. Motile spirochetes and/or motile spherules were observed in the culture fluid from all 11 patients after four weeks ( Table 2). Genital cultures from the four controls were incubated for four weeks. None of the control cultures contained visible spirochetes, and the cultures were sent for PCR testing. Genital cultures from the 11 patients were incubated for four weeks and checked weekly. Motile spirochetes were observed in the culture fluid from all 11 patients after four weeks ( Figure 1A). See Dataset, data file 1.

Most genital cultures grew very well and contained abundant spirochetes, but some blood cultures contained few spirochetes. Therefore, to better document the presence of spirochetes in culture, the culture fluid was concentrated into pellets by centrifugation ( Table 3). Spirochetes and/or spherules were detected by sectioning and special staining of paraffin blocked pellets in all the patient blood and genital cultures concentrated by centrifugation, except for blood and genital culture pellets from Patient 1 that were lost during paraffin blocking ( Table 3). Control genital culture samples were sent directly for PCR testing and were not subjected to light and darkfield microscopy.

A. Dieterle silver staining. Using standard Dieterle staining, spherules and/or spirochetal forms were visible in all patient genital cultures ( Figure 1B). Spirochetes were detected in all patient genital culture pellets except for Patient 1, whose pellet was lost during processing ( Table 3). Using the newer fixation method, spirochetes and sperm cells were visible in semen samples and showed distinct morphology ( Figure 1C). Sperm cells are known to stain with silver stains ( Pathak et al., 1979; Schmid et al., 1983). Sperm cells were seen in all semen samples except for Patients 2 and 6, who had vasectomies (data not shown). Since control genital cultures had no visible spirochetes, the control samples were sent directly for PCR testing and were not subjected to Dieterle silver staining. See Dataset, data file 2.

B. Anti-Bb immunostaining.

I. Culture fluid – University of New Haven

Genital culture fluid from Patient 1 was fixed on a SuperFrost™ Plus microscope slide and was stained with FITC-labelled polyclonal anti-Bb antibody. Staining was strongly positive, revealing well-defined spirochetes morphologically consistent with Bb ( Figure 2A). The polyclonal antibody was not reactive to T. denticola (data not shown).

II. Culture pellets – McClain laboratories

Anti-Bb immunostaining was positive for all genital cultures except for Patient 1, whose pellet was lost during processing ( Table 4). Immunostaining revealed both spiral and globular Bb forms ( Figure 2B). Since control genital cultures had no visible spirochetes, the control samples were sent directly for PCR testing and were not subjected to immunostaining. See Dataset, data file 3.

Hybridization with the Fla B probe was positive for genital culture pellets from Patients 2–9 ( Table 4). The culture pellet from Patient 1 was lost during processing. The molecular probe showed intense staining in vaginal secretions and less intense staining in semen samples ( Figure 3A and 3B). See Dataset, data file 4.

A. Australian Biologics. Borrelia 16S rRNA sequence was not detected by real-time PCR in any of the control genital culture pellets. In contrast, Borrelia 16S rRNA sequence was detected in genital culture pellets from 11 of 13 patients ( Table 5A). Patient 2 had equivocal test results and Patient 3 had negative test results in seminal cultures. See Dataset, data file 5. Real-time PCR failed to detect treponemal gene sequences in any of the control or patient genital culture pellets. See Dataset, data file 5a. The 16S rRNA isolates from six patients were sequenced and subjected to BLAST analysis (see below).

B. University of New Haven. PCR testing using the TaqMan assay for Borrelia 16S rRNA sequence was positive in blood culture pellets from seven of nine patients tested ( Table 5B). Patients 1 and 5 had negative results in blood culture pellets using the TaqMan assay, but both were positive by nested PCR for the pyrG gene. In addition, nested PCR targeting the fla gene was performed on blood culture pellets from Patients 2, 3 and 4, and nested PCR targeting the 16S rRNA gene was performed on the blood culture pellet from Patient 6. The samples were positive, and sequencing revealed 99–100% homology with Bb sensu stricto strain B-31 ( Table 5B). See Dataset, data file 7.

PCR testing using the TaqMan assay for Borrelia 16S rRNA sequence was negative in all four control genital culture pellets, and nested PCR targeting the pyrG and fla genes was negative in all four control samples, confirming the results of the TaqMan assay ( Table 5B). In contrast, eight of nine patients were positive for TaqMan 16S rRNA sequence in the genital culture pellets. Patient 6 was negative using the TaqMan assay for 16S rRNA sequence but positive using nested PCR targeting a different portion of the 16S rRNA gene ( Table 5B). Nested PCR targeting the fla gene (Patient 3) and the 16S rRNA gene (Patients 3 and 7) was also performed on genital culture pellets and was positive in those patients, confirming the results of the TaqMan assay. Patient 12 had positive PCR targeting the pyrG gene with confirmatory sequencing (see below).

PCR isolates of the vaginal culture from Patient 1 (Australian Biologics) and the seminal culture from Patient 3 (University of New Haven) were subjected to Sanger sequencing and BLAST analysis and showed 97–99% homology with Bb sensu stricto strain B-31 ( Table 5A and Table 5B). See Datasets, data files 6 and 7. PCR isolates of blood cultures from Patients 2, 3, 4 and 6 were subjected to Sanger sequencing and BLAST analysis at University of New Haven and showed 99–100% homology with Bb sensu stricto strain B-31 ( Table 5B). See Dataset, data file 7.

PCR isolates of genital cultures from three couples having unprotected sex (Patients 6-7, 10-11 and 12-13) were subjected to Sanger sequencing and BLAST analysis. Patients 6, 7, 10, 11 and 13 had sequencing done at Australian Biologics, while Patient 12 had sequencing done at University of New Haven. Sequencing revealed that the first and third couples had Borrelia strains that matched Bb sensu stricto strain B-31 ( Table 6). In contrast, the second couple had PCR sequences that matched B. hermsii strain YOR. Thus the Borrelia strain shared by this couple differed significantly from the strains identified in the other couples. See Dataset, data file 6.