|Home | About | Journals | Submit | Contact Us | Français|
PCR targeting the 16S-23S rRNA gene intergenic transcribed spacer (ITS) region has been proposed as a rapid and reliable method for the detection of Bartonella species DNA in clinical samples. Because of variation in ITS sequences among Bartonella species, a single PCR amplification can be used to detect different species within this genus. Therefore, by targeting the ITS region, multiple PCRs or additional sample-processing steps beyond the primary amplification can be avoided when attempting to achieve molecular diagnostic detection of Bartonella species. Although PCR amplification targeting this region is considered highly sensitive, amplification specificity obviously depends on primer design. We report evidence of nonspecific PCR amplification of Mesorhizobium species with previously published primers that were designed to amplify the Bartonella consensus ITS region. Use of these or other, less species-specific, primers could lead to a false-positive diagnostic test result when evaluating clinical samples. We also report the presence of Mesorhizobium species DNA as a contaminant in molecular-grade water, a series of homologous sequences in the ITS region that are common to Bartonella and Mesorhizobium species, the amplification of Mesorhizobium DNA with unpublished primers designed in our laboratory targeting the ITS region, and the subsequent design of unambiguous ITS primers that avoid nonspecific amplification of Mesorhizobium species. Our results define some potential limitations associated with the molecular detection of Bartonella species in patient samples and indicate that primer specificity is of critical importance if the ITS region is used as a diagnostic target for detection of Bartonella species.
Bacteria of the genus Bartonella (alpha subdivision of the class Proteobacteria [α-proteobacteria]) are fastidious, gram-negative, aerobic bacilli with more than 17 described species (12, 23, 24, 25, 37). Because of their zoonotic potential; transmission by a wide range of insect vectors, including sand flies, lice, fleas, and ticks; and ability to persistently infect mammalian reservoir hosts, Bartonella bacteria are considered emerging pathogens (1, 6, 8, 10, 48). Eight species of this genus have been isolated from humans and are the causative agents of Carrion's disease, Oroya fever, verruga peruana (Bartonella bacilliformis) (6), trench fever (B. quintana) (9, 40), endocarditis (B. elizabethae, B. henselae, B. quintana, B. vinsonii subsp. arupensis, B. vinsonii subsp. berkhoffii, and B. washoensis) (13, 15, 16, 20), bacillary angiomatosis in immunocompromised patients (B. quintana and B. henselae) (12, 28, 51), and cat scratch disease (B. henselae and B. clarridgeiae) (3, 30, 42). In addition, B. grahamii has been diagnosed by PCR as the causative agent of neuroretinitis. (27) Other Bartonella species, including B. alsatica, B. doshiae, B. grahamii, B. henselae, B. koehlerae; B. peromysci, B. talpae, B. taylorii, B. tribocorum, and B. bovis, have been isolated from animals including cats, dogs, deer, cattle, and lions, among others (10, 14, 16, 17, 19, 21, 22, 29, 31, 41, 48, 52).
Because of their fastidious nature and their characteristically slow growth (which can require up to 45 days of incubation), microbiological diagnosis of diseases caused by Bartonella species by various isolation procedures is difficult or, in many cases, impossible (38, 43, 46, 49, 50, 51). Because of the rapidly increasing number of Bartonella species implicated as animal or human pathogens, limitations associated with culturing these organisms from patient samples, and the undefined or inconsistent serological cross-reactivity among Bartonella species, molecular detection of Bartonella DNA in patient samples provides a potentially attractive alternative for diagnostic purposes. During the last 5 years, an increasing number of research laboratories have focused on the development of new molecular-level diagnostic methods for the rapid identification and differentiation of Bartonella species. Molecular approaches have included analysis by (i) sequencing or restriction fragment length polymorphism analysis of PCR-amplified genes, such as the 16S rRNA gene, the 16S-23S rRNA intergenic transcribed spacer (ITS) region, the citrate synthase gene, the riboflavin synthase gene, the groEL gene, or the RNA polymerase beta subunit gene (2, 4, 5, 25, 33, 35, 36, 38, 39, 44, 45, 47, 53, 55); (ii) species-specific amplification of the ftsZ gene (18, 26, 54); and (iii) direct species subtyping by PCR amplification with genus-specific primers based on the ITS region (7, 23, 24). Among these approaches, species subtyping with ITS-based, genus-specific primers has proven to be one of the most practical methods for diagnostic detection of Bartonella species. ITS subtyping relies on the variation in species-specific amplicon size obtained by a single PCR, without requiring multiple PCR amplifications or additional sample-processing steps beyond the primary PCR amplification.
Although PCR amplification of the ITS region is a rapid method for Bartonella DNA detection and subtyping, specificity depends on the primer design. Nonspecific amplification of contaminant bacterial DNA could result in a misdiagnosis of bartonellosis, particularly if critical attention was not given to amplicon size or location on the gel. Misdiagnosis could cause the clinician to select an ineffective antibiotic or treat for an inappropriate duration or could result in a delayed diagnosis of a more serious, noninfectious disease. As infection with Bartonella species can accompany serious illnesses such as endocarditis, granulomatous hepatitis, osteolytic lesions, and neuroretinitis, failure to arrive at an accurate microbiological diagnosis of bartonellosis could also adversely influence patient care. We report here consensus sequences that are common to the 16S-23S ITS regions of Bartonella and Mesorhizobium spp. and may result in the potential amplification of Mesorhizobium DNA from patient samples undergoing testing for Bartonella infection. We also document, by sequencing, nonspecific amplification of Mesorhizobium species with previously published primers and primers designed in our laboratory targeting the Bartonella ITS region. A specific set of Bartonella ITS primers that avoid nonspecific amplification of Mesorhizobium DNA are described.
Pure cultures of B. bovis (NCSU99-B01) isolated at the North Carolina State University College of Veterinary Medicine and B. clarridgeiae (ATCC 700095), B. elizabethae (ATCC 49927), B. henselae Houston-1 (ATCC 49882), B. quintana Fuller (ATCC 51694), and B. vinsonii subsp. berkhoffii (ATCC 51672) obtained from the American Type Culture Collection (Manassas, Va.) were used for DNA extraction and PCR amplification of the Bartonella species ITS region.
DNAs from the above-described pure cultures of B. bovis, B. clarridgeiae, B. elizabethae, B. henselae Houston-1, B. quintana Fuller, and B. vinsonii subsp. berkhoffii were prepared by using the QIAamp DNA Mini Kit (QIAGEN Inc., Valencia, Calif.). After extraction, DNA concentration and purity were measured with an absorbance ratio of 260 to 280 nm. Each DNA extraction was then diluted to 1.0 pg/μl with 1× Tris-EDTA buffer, pH 7.4. This concentration of DNA was used as the standard for all PCR analysis.
In order to examine ITS region-based primer specificity for the detection of Bartonella species, sequences from the GenBank database were aligned with AlignX software (InforMax, Inc., Bethesda, Md.). The ITS regions of the following Bartonella species were included in the alignments: B. henselae strains URBHLIE-9 (accession no. AF312496), URBHLLY-8 (accession no. AF312495), 90-615 (AF369528), CAL-1 (accession no. AF369527), Fizz (accession no. AF365226), and Houston-1 (accession no. L35101); B. quintana strains URQBTBAAH-1 (accession no. AF368396), Fuller (accession no. L35100), URBQMTF-95 (accession no. AF369482), Oklahoma (accession no. AF368391), SH-PERM (accession no. AF368392), URBQMLY-4 (accession no. AF368393), and URBQPIEH-2 (accession no. AF368394); B. taylorii strains M6 (accession no. AF269784) and wm9 (accession no. AJ269788); and B. vinsonii subsp. vinsonii Baker (accession no. L35102), berkhoffii (accession no. AF312503), and arupensis (accession no. AF442952). The ITS sequence of B. henselae strain Houston-1 (GenBank accession no. L35101) was used as a reference for primer numbering identification.
After the alignment of Bartonella ITS sequences, two consensus sequence regions were targeted for amplification. For the first region, oligonucleotides H56s (5′ GGG GAA CCT GTG GCT GGA TCA C 3′) and H493as (5′ TGA ACC TCC GAC CTC ACG CTT ATC 3′) were synthesized as sense and antisense primers, respectively. The second region targeted for amplification used oligonucleotides 321s (5′ AGA TGA TGA TCC CAA GCC TTC TGG 3′) and 983as (5′ TGT TCT YAC AAC AAT GAT GAT G 3′) as sense and antisense primers, respectively (see Results and Discussion for details). All primers were synthesized by IDT Integrated DNA Technology (Coralville, Iowa).
PCR amplification was performed with a reaction mixture with a 25-μl final volume containing 17 μl of molecular-grade water (Epicentre), 0.5 μl of a 10 mM deoxynucleoside triphosphate mixture, 2.5 μl of 10× PCR buffer, 2.5 μl of 25 mM MgCl2, and 0.7 U of AmpliTaq Gold DNA polymerase, all from PE Applied Biosystems (Foster City, Calif.). The reaction mixture was completed by adding 0.25 μl of the forward and reverse primers (each at 30 μM) and 2 μl of either 1.0 pg of DNA from each Bartonella species tested per μl, 1.0 pg of Escherichia coli DNA (as an outside group PCR negative control) per μl, or water (as a PCR negative control).
Amplification of the first ITS consensus region was performed under the following conditions: a hot-start cycle of 95°C for 5 min, followed by 45 cycles of denaturing at 95°C for 30 s, annealing at 60°C for 30 s, and extension at 72°C for 45 s. Amplification was completed by an additional cycle of 72°C for 10 min to allow complete extension of the PCR products. Amplification of the second ITS consensus region was performed with the following conditions: a single hot-start cycle of 95°C for 5 min, followed by 45 cycles of denaturing at 94°C for 45 s, annealing at 54°C for 45 s, and extension at 72°C for 45 s. As described above, amplification was completed by an additional cycle of 72°C for 10 min to allow complete extension of the PCR products. All amplification products were visualized by ethidium bromide staining after electrophoresis through a 2% agarose gel. Amplicon sizes were determined by comparison with the molecular weight Hyladder 1-kbp marker from Denville Scientific Inc.
Bartonella ITS PCR amplification products obtained with primers H56s and H493as derived from PCR negative water controls or from EDTA-anticoagulated dog or bear blood samples were cloned into the plasmid pGEM-T Easy vector system (Promega, Madison, Wis.). By white color screening, recombinant colonies were selected; plasmids, prepared with QIAGEN′s plasmid miniprep, were chosen on the basis of the correct insert size. For sequencing, 1.5 pmol each of IRD-labeled M13 forward and reverse primers was used as recommended by the LI-COR Company (Lincoln, Nebr.). Both strands of recombinant plasmid (200 to 400 fmol) were sequenced with the SequiTherm Excel II Long Read DNA sequencing kit (LC, Epicenter Technologies). The sequencing reaction mixtures were analyzed by polyacrylamide gel electrophoresis on a LI-COR 4200 automated DNA sequencer, and alignment with BlastX was performed in order to identify bacteria at the genus and species levels. Each PCR amplicon was cloned and sequenced in triplicate.
After alignment of the ITS region sequences of the above Bartonella species, several highly conserved regions were found (homology A in Fig. Fig.1).1). The most noticeable homologous regions (as relative positions) were located at positions 1 to 100, 280 to 420, 590 to 700, 980 to 1120, and 1150 to 1450. An initial set of primers, described as H56s and H493as, were designed and synthesized for the amplification of the two hypervariable ITS regions located between the first and third homologous regions.
PCR amplification of DNAs from B. bovis, B. clarridgeiae, B. elizabethae, B. henselae Houston-1, B. quintana Fuller, and B. vinsonii subsp. berkhoffii with primers H56s and H493as resulted in a single amplicon for each Bartonella species (Fig. (Fig.2).2). PCR product sizes ranged from 419 bp (B. bovis) to 565 bp (B. elizabethae) and were dependent on the species. The observed amplicon sizes were in accordance with the theoretical ITS amplicon sizes estimated on the basis of the data available from the GenBank database (Table (Table1).1). These preliminary results suggested that this region could facilitate rapid detection of Bartonella species in patient samples; however, marginal differences in some amplicon sizes would not facilitate definitive identification of all Bartonella species (such as differences between B. henselae and B. koehlerae or between B. quintana and B. taylorii). Use of E. coli template DNA produced no amplification under the same PCR conditions. Unfortunately, when the PCRs were conducted with water from different sources (injection grade or molecular grade) as the negative control template, nonspecific amplification products were obtained (Fig. (Fig.2).2). Negative control amplicons were approximately 420 bp in size, which is between the amplification product sizes of B. bovis and B. clarridgeiae. Similarly, a 420-bp amplicon was obtained when DNA extracted from several dog and bear clinical samples (EDTA-anticoagulated blood) was amplified with primers H56s and H493as (results not shown).
In order to establish the identity of the nonspecific amplicons, PCR products obtained from negative control water sources and from animal DNA samples were cloned and sequenced as described above. Sequencing and alignment analysis with BlastN from GenBank of all 420-bp amplicons matched the ITS region of Mesorhizobium loti with 99.9% homology. These results suggested that the DNA of this α-proteobacterium was present as a contaminant in some component of the PCR mixture (molecular-grade water, DNA extraction buffers, or the columns).
The same nonspecific PCR amplification was obtained when previously published primers were tested with the same molecular- and non-molecular-grade water samples. In all instances, M. loti was amplified (Fig. (Fig.3)3) when PCRs targeted the amplification of the ITS region between positions 384 and 638 with oligonucleotides J1 (5′ YCC TTC GTT TCT CTT TCT TCA 3′) and J2 (5′ AAC CAC TGA GCT ACA AGC C 3′) as sense and antisense primers, respectively, as described by Jensen et al. (24).
To further delineate similarities between Bartonella and Mesorhizobium sequences, the ITS regions of B. henselae, B. quintana, and B. vinsonii subspecies (as described above) were aligned with the ITS sequences of M. loti (accession no. AP003001) and M. mediterraneum (accession no. AF345262). Sequence alignment analysis showed several homologous ITS regions among the members of the two genera (Fig. (Fig.1,1, homology B). Sequence analysis also indicated that both sets of tested primers (H56s-H493as and J1-J2) were located within these homologous regions.
A new set of Bartonella ITS primers was designed to avoid nonspecific amplification of Mesorhizobium species. Oligonucleotide 321s and 983as, as described in Materials and Methods, were designed for amplification of the hypervariable ITS region located between positions 384 and 1172 (Fig. (Fig.1,1, homologies A and B). On the basis of alignment with GenBank sequences, these new primers did not amplify DNA from other bacteria but did amplify DNAs from all of the Bartonella species previously tested. Amplicon sizes were species dependent, ranging from 704 bp for B. vinsonii subsp. berkhoffii to 453 bp for B. bovis (Fig. (Fig.4).4). All Bartonella species PCR products matched the theoretical amplicon sizes estimated from sequences derived from GenBank (Table (Table2).2). The new primers (321s and 983as) did not amplify Mesorhizobium species from molecular-grade water or non-molecular-grade (for injection) water (including water that previously resulted in nonspecific amplification of Mesorhizobium species) or when DNAs extracted from blood samples from non-Bartonella-seroreactive dogs were used as templates. In addition, no amplification was obtained when DNAs extracted from several Ehrlichia species and several Rickettsia species (closely related α-proteobacteria) were tested (results not shown).
Our laboratory contributed to the validation of the 16S-23S ITS Bartonella primers reported by Jensen et al. (24). On the basis of the results of this study, the use of those primers could result in nonspecific amplification of Mesorhizobium species from patient samples. Mesorhizobium species, including M. loti, are common environmental bacteria and can be found in soil, plants, and water (11, 32). Therefore, the potential detection of Mesorhizobium DNA in non-molecular-grade water is not a surprising finding. As illustrated in this study, contamination of some sources of molecular-grade water is also possible. Although not specifically examined in this study, a number of other published Bartonella ITS primers, on the basis of sequence alignment, could in fact amplify Mesorhizobium species DNA if present in the test sample or the PCR mixture. In addition to the primers described by Jensen et al. (24), Houpikian and Raoult (23) and Birtles et al. (7) reported the use of primers 16SF (or QHVE1) (5′ AGA GGC AGG CAA CCA CGG TA 3′) and 23S1 (or QHEV2) (5′ GCC AAG GCA TCC ACC 3′) for specific amplification of the 16S-23S rRNA intergenic region of the genus Bartonella. Sequence alignment indicates that all of these primer sets, as well as the primer set initially developed in our laboratory (H56s-H493as), contain 100% consensus sequence homology within the M. loti ITS region. Another set of primers reported by Houpikian and Raoult (23), designed to target B. bacilliformis and B. quintana in human blood samples, also has a high level of sequence homology with the same M. loti ITS region.
After comparison of homologous sequences between the Bartonella and Mesorhizobium ITS regions, a new set of primers was designed to specifically amplify Bartonella species but not Mesorhizobium species. The new primers (321s and 983as) did not amplify other closely related α-proteobacteria, including Ehrlichia or Rickettsia species or M. loti, from blood samples obtained from healthy or sick dogs that lacked Bartonella species antibodies. In addition, these primers did not amplify Mesorhizobium DNA from several sources of water used as PCR negative controls that were previously shown to contain M. loti DNA by amplification, cloning, and sequencing.
On the basis of the results of this study, the careful selection and design of ITS primers for the molecular detection of Bartonella species, when used for diagnostic or research purposes, are crucial to avoid false-positive test results. A large number of previously described molecule-based diagnostic tests targeting the ITS region could lead to nonspecific amplification of the DNA of a species other than a Bartonella species. The contamination of water-containing solutions with extraneous bacterial DNA, including DNA extraction buffers, molecular-grade water, or enzyme preparations seems to be a relatively common molecular diagnostic problem (11, 32, 34) that poses a risk of misdiagnosis and the selection of inappropriate treatments. As illustrated recently for Chlamydia pneumoniae, nonspecific amplification as a result of extraneous bacterial DNA contamination could lead to the incorrect identification of a putative causative agent (34). Although of obvious importance for the accurate interpretation of research studies, nonspecific amplification of contaminant bacterial DNA is of critical importance when molecular diagnostic test results are used to direct individual treatment decisions for animals or for human patients (34). The inadvertent amplification of Mesorhizobium DNA when using molecular-grade water designed for PCRs could lead to erroneous research or clinical interpretations. In conclusion, the results of this study should assist microbiologists attempting to improve the molecular diagnostic detection of Bartonella species in patient samples and other investigators using the 16S-23S ITS region as a molecular target in epidemiological or pathological studies involving animal or human disease. We suggest the use of hot-start DNA polymerases, the inclusion of positive and negative controls on each agarose gel, and the cautious interpretation of test results that are based on agarose gel electrophoresis or when non-sequence-specific dyes (such as SYBR green) are used for DNA amplification by real-time PCR.
The State of North Carolina supported this research.