2.1. Detection of “Lyme Borreliae” in Archived Post-Treatment Sera from Lyme Disease Patients
Under a Material Transfer Agreement with the CDC of the United States, the authors received from the latter agency 32 blind-coded serum samples (100 μL each), including 12 serum samples collected from patients with clinically suspect Lyme disease according to the CDC criteria and had been treated with antibiotics for the infection, and 20 control serum samples collected from patients, two diagnosed with fibromyalgia, two with rheumatoid arthritis, two with multiple sclerosis, two with infectious mononucleosis, two with syphilis and two with severe periodontitis; and from healthy persons, four living in Lyme disease-endemic regions, and four living in non-endemic regions. These CDC samples were serology references, which had been tested by a CDC-approved reference laboratory and were known to contain the standard Lyme disease antibody bands generated by an approved 2-tier serology test kit in the positive samples. They are intended to be used for evaluation of the accuracy of new diagnostic tests for Lyme disease. The samples were shipped to the first author’s laboratory in frozen state over dry ice in a thermo-insulated container by overnight delivery from the Centers for Disease Control and Prevention, Fort Collins, CO, USA.
The serum samples were allowed to thaw at room temperature. The bacteria in the serum were pelleted by centrifugation and the microbial DNA in the pellet was extracted with hot ammonium hydroxide and precipitated in ethanol. The DNA was finally dissolved in 100 μL TE buffer. Two 3 μL aliquots of each DNA extract were used to initiate two same-nested PCRs [17
] for each sample in duplicate. The same-nested PCR consisted of a primary PCR and a nested PCR in tandem with one identical pair of M1 and M2 primers for both. The M1 and M2 primers were “genus-specific” primers designed to amplify a highly conserved segment with hypervarable regions of the borrelial16S ribosomal RNA gene DNA (16S rDNA) of all commonly known borrelia species found in the GenBank. The nested PCR products were subjected to agarose gel electrophoresis. The 357–358 bp target amplicon was used as the template for automated two-directional Sanger sequencing with the M1 and M2 nucleotide as the sequencing primer for each direction. A minimum 150-base inter-primer segment of the signature DNA sequence excised from the computer-generated base-calling electropherogram was submitted to the GenBank for BLAST (Basic Local Alignment Search Tool) alignment analysis. A 100% identities (ID) match with a standard sequence on the GenBank report was required for validation of the molecular diagnosis of a borrelial species.
After the testing results were reported to the CDC, the blinded serum samples were decoded. Matching the PCR results with the CDC decoding data (Table 1
) showed that among the 32 samples tested only one of the 3 μL duplicate aliquots of the DNA extract from the serum of patient #9 generated a positive M1/M2 PCR amplion.
The #9 patient was diagnosed with “neurologic Lyme disease” and had been treated before the serum sample was drawn. Direct DNA sequencing of the nested PCR amplicon confirmed that the sequence of the amplicon is that of a novel borrelia in the relapsing fever group (see below). Based on previously published nested PCR data [18
], this result indicated that the number of residual borrelial bacteria or bacterial fragments containing chromosomal DNA in the positive serum sample was <30 per 100 μL.
Since >95% of the B. burgdorferi
spirochetes suspended in whole blood are usually lost during preparation of the serum samples [21
] for serology testing, little efforts have been made to detect infectious agents in the split serum sample when the 2-tier serology test is performed for Lyme disease diagnosis in the United States. In one study reported from Poland, Niścigorska et al.
found no DNA of B. burgdorferi
sensu lato in 52 split serum samples of forestry workers although 32 of them showed serology evidence of B. burgdorferi
senso lato infection [22
2.2. Sequencing of a Highly Conserved Segment of 16S rDNA of a Novel Borrelia
A small amount of the 358-bp M1/M2 amplicon from the #9 same-nested PCR tube was transferred directly by a micro-glass rod into 20 μL of automated Sanger reaction mixture [9
] containing the M1 nucleotide or the M2 nucleotide as the sequencing primer. The computer-generated two-directional DNA sequencing base-calling electropherograms, including the primer sites, are shown in Figure 1A,B
Connecting the two sequences (Figure 1A,B
) after all the complementary bases were converted to those for a 5′–3′ reading resulted in a composite inter-primer segment of 316-base 16S rDNA sequence. Submission of this sequence to the GenBank for BLAST alignment generated a returned report. A copy of the report is presented as Figure 1C
This GenBank report indicated that there is no 100% ID match for the submitted unique DNA sequence with any of the nucleotide sequences stored in the GenBank databases. The closest is a 314/316 bases ID match (Figure 1C
) with B. coriaceae
(AF210135) and a 312/316 bases ID match with B. hermsii
(CP004146), B. anserina
(CP005829), B. parkeri
(CP005851) and B. turicatae
(AY604974), a group of relapsing fever borreliae. Attempts to further characterize this novel isolate failed due to the limited number of microbes in the serum sample.
As shown in the above GenBank report (Figure 1C
), the unique 316 bases 16S rDNA segment of the novel borrelia shares a common sequence with B. coriaceae
with two discordant bases at positions 956 and 964 where the two nucleotide bases “A” have been replaced by two nucleotides “G” (highlighted in red in the above BLAST alignment report). For comparison, the chromosomal DNA of a standard culture of B. coriaceae
, strain Co53 (ATCC 43381) was extracted and amplified with the M1/M2 same-nested PCR. The two-directional sequencing electropherograms of the corresponding segment of the B. coriaceae
16S rDNA are presented in Figure 2
Based on animal studies, B. burgdorferi
is highly susceptible to the antibiotics chosen for Lyme disease treatment, and is invariably eliminated from the blood after a course of antibiotic treatment [23
]. Therefore, the finding of “Lyme disease bacteria” in the post-treatment serum sample of a patient with neurologic Lyme disease was a surprise. It was more surprising when a novel borrelial 16S rDNA sequence was identified instead of that of B. burgdorferi
as the residual bacteria in the serum.
PCR amplification of 16S rDNA followed by direct DNA sequencing has been used for construction of the phylogenetic tree of the borrelia species [13
], thus is highly reliable for molecular diagnosis of borrelial infectious agents in clinical samples. However, the usefulness of 16S rDNA sequencing as a tool in microbial identification is dependent upon two key elements, deposition of complete unambiguous nucleotide sequences into public or private databases and applying the correct “label” to each sequence [25
]. A search of the DNA sequence databases of the GenBank confirmed that the nucleotides at positions 956 and 964 of the 16S ribosomal RNA gene of all known borrelia species are always occupied by the adenine base “A”, using the B. coriaceae
16S rDNA (GenBank locus #AF210135) as the positions reference. In order to rule out the possibility that the guanine bases “G” at positions 956 and 964 observed in Figure 1B
(underlined) were not a result of polymerase-induced mutations which are invariably random, more than 4 same-nested PCRs, each followed by a two-directional sequencing of the PCR amplicon, were performed on the DNA extracts from sample #9 (Table 1
) and from the standard B. coriaceae
culture to confirm the reproducibility of the sequences represented in Figures 1
. All sequences generated from sample #9 and from the standard culture were identical to those illustrated in Figures 1
is a species known to infect the soft tick Ornithodoros coriaceus
, classified in the relapsing fever group, and not known to be a human pathogen [26
]. Whether this novel isolate represents a true mutant of a wild-type B. coriaceae
or a new borrelia species of the highly heterogeneous relapsing fever group would have to await further studies.
Since the residual bacteria or bacterial remnants of this novel isolate were found in the serum of a patient who had been diagnosed with and treated for “neurologic Lyme disease”, and since the concomitant 2-tier serology test performed on the split sample showed a classic Lyme disease antibody band pattern (Table 1
), it is possible that patient #9 might have a mixed infection by a B. burgdorferi
and a novel relapsing fever borrelia of uncertain significance which was more resistant to the standard regimen of antibiotic treatment than the B. burgdorferi
co-infectant. Alternatively, this novel bacterium as a sole infectious agent of this patient’s neuroborreliosis might have some common epitopes with the B. burgdorferi
species, causing a positive Lyme disease 2-tier serology test result in the host’s serum.
2.3. Detection of “Lyme Borreliae” in Archived Pre-Treatment Sera from Lyme Disease Patients
Under another Material Transfer Agreement, the authors received from the CDC 20 blind-coded serum samples (100 μL each), including an undisclosed number of serum samples collected from patients who were diagnosed with clinically suspect Lyme disease according to the CDC criteria, but had not been treated with antibiotics for the infection at the time when the serum samples were collected for the 2-tier serology test.
This second batch of 20 serum samples was shipped to the first’s author’s laboratory, and the DNA of the residual bacteria or bacterial fragments in the sera was extracted and subjected to M1/M2 same-nested PCR as described above. The 357–358 bp nested PCR amplicons detected were sequenced for validation. After the testing results were submitted to the CDC, the blinded samples were decoded. Matching the PCR results with the CDC decoding data (Table 2
) showed that only 3 samples of the sera taken from 20 patients all with a diagnosis of clinically suspect Lyme disease were positive for infectious agents, including one isolate of B. miyamotoi
(#39) and two isolates of B. burgdorferi
sensu lato (#45 and #50). For each PCR-positive serum sample, the 16S rDNA was only amplified successfully in one of the two duplicate PCRs, indicating that the number of borrelial bacteria or bacterial fragments in the 100 μL serum sample was <30.
Since DNA sequencing-based tests for Lyme disease have not been widely used in laboratory medicine in the United States, the direct Sanger sequencing data are presented below as examples for its application in the diagnosis of B. miyamotoi and B. burgdorferi in clinical materials.
2.4. DNA Sequencing-Based Diagnosis of B. miyamotoi and B. burgdorferi
The 358 bp M1/M2 primer-terminated PCR amplicon derived from serum sample #39 (Table 2
) was subjected to two-directional DNA sequencing as described above. A segment of the 16S rDNA signature sequence (Figure 3
) was submitted to the GenBank for BLAST sequence alignment analysis, as an example for routinely validating the molecular diagnosis of a B. miyamotoi
in a clinical laboratory.
It is of interest to note that this patient had a diagnosis of clinically suspect early Lyme disease with erythema migrans
(EM) and that there were 2 of 3 specific bands in the IgM immunoblot although the 2-tier serology test was interpreted as negative because there were a low EIA value and only 2 bands in the IgG immunoblot (Table 2
). A DNA sequencing of an M1/M2 primer-terminated PCR amplicon would have confirmed that patient #39 actually had a B. miyamotoi
infection with a skin rash, similar to some of the patients reported in the Russian series [13
The 357 bp M1/M2 primer-terminated PCR amplicon derived from serum sample #45 (Table 2
) was subjected to a two-directional DNA sequencing as described above. A segment of the 16S rDNA signature sequence (Figure 4
) was submitted to the GenBank for BLAST sequence alignment analysis, as an example for routinely validating the molecular diagnosis of B. burgdorferi
in a clinical laboratory.
Although patient #45 (Table 2
) had a diagnosis of clinically suspect early Lyme disease with EM, the 2-tier serology test was not supportive of the clinical diagnosis due to a low EIA value and only one band observed in the IgG immunoblot with no bands in the IgM immunoblot. For patients similar to case #45, the finding of B. burgdorferi
16S rDNA validated by DNA sequencing would be of great value in confirming the clinical diagnosis.
Serum sample #50 was confirmed by PCR and DNA sequencing to be positive for B. burgdorferi
, similar to serum sample #45. Patient #50 was diagnosed with clinically suspect early Lyme disease with a high EIA value and 2 of 3 specific bands in the IgM immunoblot, but only one IgG band. Although the CDC reference laboratory interpreted the 2-tier serology test results in sample #50 as positive, some practitioners might consider these combinations of bands to be ambiguous or borderline, based on the CDC guidelines which require demonstration of five of the 10 bands in an IgG immunoblot for a positive 2-tier serology test [10
]. Patients with a set of 2-tier serology test results like those in case #50 may or may not be diagnosed as having Lyme disease if the clinical presentations are not typical of “Lyme disease”. A reliable test for the infectious agents in EDTA-anticoagulated whole blood samples would be of great help in confirming the clinical diagnosis in these cases. The CDC has not maintained a panel of standard borrelia-positive whole blood samples taken from Lyme disease patients for diagnostic methodology evaluation.