OUP user menu

Comparative analysis of Borrelia isolates from southeastern USA based on randomly amplified polymorphic DNA fingerprint and 16S ribosomal gene sequence analyses

Tao Lin, James H. Oliver Jr., Lihui Gao
DOI: http://dx.doi.org/10.1016/S0378-1097(03)00763-8 249-257 First published online: 1 November 2003

Abstract

Fifty-three southern USA Borrelia isolates were characterized using randomly amplified polymorphic DNA fingerprinting analysis (RAPD). Twenty-nine types were recognized among 37 B. andersonii strains, seven types among eight B. bissettii strains, and seven types among seven B. burgdorferi sensu stricto strains. Strain TXW-1 formed a separate RAPD type. Nearly complete sequences of the rrs genes from 17 representative southern Borrelia were determined. The similarity values were found to be 96–100% within the B. burgdorferi sensu lato (s.l.) complex, 94–99% among the relapsing fever borreliae, and 93–99% between the two complexes. Phylogenetic analysis indicated that all the Borrelia strains we analyzed could be divided into two parts: the B. burgdorferi s.l. complex and the relapsing fever borreliae complex. TXW-1 segregated with the North American relapsing fever borreliae and formed a separate subbranch.

Keywords
  • Lyme disease
  • Borrelia
  • Randomly amplified polymorphic DNA fingerprinting analysis
  • 16S ribosomal gene

1 Introduction

Lyme borreliosis is the most prevalent tick-borne disease in temperate regions throughout the world. It is caused by certain genospecies of the Borrelia burgdorferi sensu lato (s.l.) complex and transmitted by several species of the Ixodes ricinus species complex. More than 80% of Lyme borreliosis cases in the USA occur along the northeastern and mid-Atlantic seaboards [1]. Different clinical symptoms of Lyme borreliosis have been associated with distinct genospecies of B. burgdorferi s.l. [2].

Relapsing fever is another medically important tick-borne borreliosis and is characterized by occurrence of one or more spells of fever after the initial febrile attack. There are two forms of relapsing fever: louse-borne or epidemic relapsing fever caused by B. recurrentis and transmitted by the human body louse, and tick-borne or endemic relapsing fever caused by at least 16 distinctive Borrelia species and transmitted by several species of argasid ticks. Some of the latter Borrelia exhibit complete specificity for their tick vectors. Approximately 20 species of relapsing fever borreliae have been identified to date [3,4].

The genus Borrelia is comprised of two species complexes, B. burgdorferi complex and relapsing fever borreliae, based on genetic differences, specificity of the spirochete–vector relationship, and the type of disease they cause. This genus can be distinguished from other spirochetal phylogenetic groups by base signature analysis of rrs [5].

At least 11 genospecies have been described in the B. burgdorferi s.l. complex. Three of them, B. burgdorferi sensu stricto (s.s.), B. garinii, and B. afzelii, have been isolated from patients. Genetic analysis of ospC and clinical studies indicate that the major ospC groups found in human erythema migrans (EM) lesions are only a subset of the ospC groups found in ticks [6,7] and there is only a minority (10/58) of ospC groups in the three genospecies (B. burgdorferi s.l., B. garinii, and B. afzelii) that are believed to be invasive and able to colonize and persist in organs other than skin. Strains in other groups may cause EM, but do not appear to be invasive and disseminate to other organs.

Three genospecies within the B. burgdorferi s.l. complex have been identified in the southern USA [810]. However, strains from the same genospecies sometimes are difficult to differentiate based on restriction fragment length polymorphism (RFLP) analysis of rrf-rrl intergenic spacer. Moreover, rrf-rrl intergenic spacer is only a DNA fragment, and does not reflect the whole genomic situation, nor does the intergenic spacer relate to pathogenicity of Borrelia. Randomly amplified polymorphic DNA (RAPD) fingerprinting analysis is a good genomic indicator and also may provide information related to pathogenicity. Finally, the taxonomic status of some southern strains remains uncertain and needs analysis using other molecular techniques.

RAPD has been used for molecular typing of various microorganisms [11] including B. burgdorferi [1214]. This technique uses low-stringency polymerase chain reaction (PCR) amplification with a single primer with an arbitrary sequence to generate strain-specific arrays of anonymous DNA fragments. Since RAPD is a highly discriminatory method and easy to perform, it is a powerful tool to distinguish the different Borrelia species from each other as well as to recognize Borrelia strains within each of the species.

In the present study, we use the RAPD method to identify and classify southern Borrelia between and within genospecies, explore the relationship between Borrelia and their tick vectors and mammal reservoirs, and assess the geographic diversity within different Borrelia species. Also, representative strains from each RAPD group were chosen and their rrs genes were sequenced to differentiate spirochetes from B. burgdorferi s.l. and relapsing fever borreliae.

2 Materials and methods

2.1 Spirochete isolates and culture conditions, extraction of spirochete DNA, RAPD, and RAPD fingerprinting analysis

The isolates employed in this study are described in Table 1. Most of them have been characterized previously by rrf-rrl intergenic spacer and ospC [9,10]. Several strains including B-31, SH-2-82 (B. burgdorferi s.l.), DN127, 25015 (B. bissettii), 21038 (B. andersonii), 20047 (B. garinii) were used as positive controls. All southern isolates analyzed were cloned by a range dilution and growth on 1.3% BSK-H agar plates. These strains were third passage. Cultures were incubated at 34°C for 1–2 weeks until the cell density reached about 2×106 cells ml−1.

View this table:
Table 1

Origins, genospecies, and RAPD types of Borrelia strains analyzed in this study

GenospeciesRAPD typesIsolateHostSourceLocation
CombinedAP1312831254
B. burgdorferi s.s.1A1B1C1B31I. scapularisShelter Island, New York
2A2B2C2SH-2-82I. scapularisNew York
3A3B3C3SI-1P. gossypinusBladderSapelo Island, McIntosh County, Georgia
5A5B5C5SI-3S. hispidusBladderSapelo Island, McIntosh County, Georgia
6A6B6C6SI-14I. affinis femaleDragSapelo Island, McIntosh County, Georgia
7A6B7C6SCI-2P. gossypinusEar clipSt. Catherines Island, Liberty County, Georgia
8A7B6C6SCI-4I. scapularis maleDragSt. Catherines Island, Liberty County, Georgia
4A4B4C4MI-2P. gossypinusBladder, ear clipMerritt Island, Brevard County, Florida
9A8B8C7SM-1P. gossypinusEar clipSt. Marys, Camden County, Georgia
B. bissettii10A9B9C8DN127I. pacificusCalifornia
11A10B10C825015I. scapularis larvaWhite-footed mouseDutchess County, New York
12A11B11C9MI-6S. hispidusEar clipMerritt Island, Brevard County, Florida
13A12B12C10MI-8S. hispidusEar clipMerritt Island, Brevard County, Florida
14A13B13C11MI-9P. gossypinusEar clipMerritt Island, Brevard County, Florida
15A10B14C12SCGT-8aI. minor maleWoodratGeorgetown County, South Carolina
16A9B15C13SCGT-10N. floridanaEar clipGeorgetown County, South Carolina
17A14B16C14AI-1S. hispidusBladder, ear clipAmelia Island, Nassau County, Florida
14A13B13C11FD-1S. hispidusBladder, ear clipFavor Dykes, Flagler County, Florida
48A35B33C36SCW-30hI. minor nymphCarolina wrenWedge Plantation, Charleston County, South Carolina
B. andersonii18A15B17C1521038I. dentatusNew York
19A16B18C16MOR-1I. dentatus nymphRabbitSwinton, Stoddard County, Missouri
20A17B19C16MOR-2I. dentatus nymphRabbitSwinton, Stoddard County, Missouri
21A18B20C17MOS-1bI. dentatus larvaRabbitSwinton, Stoddard County, Missouri
21A18B20C17MOS-1cI. dentatus larvaRabbitSwinton, Stoddard County, Missouri
21A18B20C17MOS-1dI. dentatus larvaRabbitSwinton, Stoddard County, Missouri
22A19B18C18MOS-1eI. dentatus larvaRabbitSwinton, Stoddard County, Missouri
23A20B20C17MOS-1fI. dentatus larvaRabbitSwinton, Stoddard County, Missouri
22A19B18C18MOS-1gI. dentatus larvaRabbitSwinton, Stoddard County, Missouri
24A21B18C19MOS-2aI. dentatus nymphRabbitSwinton, Stoddard County, Missouri
25A18B21C20MOS-2bI. dentatus nymphRabbitSwinton, Stoddard County, Missouri
24A21B18C19MOS-3aI. dentatus nymphRabbitSwinton, Stoddard County, Missouri
25A18B21C20MOS-3bI. dentatus nymphRabbitSwinton, Stoddard County, Missouri
26A18B20C21MOS-3cI. dentatus nymphRabbitSwinton, Stoddard County, Missouri
26A18B20C21MOS-4I. dentatus nymphRabbitSwinton, Stoddard County, Missouri
27A22B22C22MOD-12 I. dentatus nymphsRabbitDowd Farm, Bollinger County, Missouri
28A22B20C23MOD-34 I. dentatus larvaeRabbitDowd Farm, Bollinger County, Missouri
29A23B20C24MOD-5I. dentatus nymphRabbitDowd Farm, Bollinger County, Missouri
30A24B20C25MOK-1bI. dentatus nymphRabbitKoch Farm, Bollinger County, Missouri
31A25B20C26MOK-1cI. dentatus nymphRabbitKoch Farm, Bollinger County, Missouri
32A26B23C26MOK-1dI. dentatus nymphRabbitKoch Farm, Bollinger County, Missouri
33A27B24C27MOK-1eI. dentatus nymphRabbitKoch Farm, Bollinger County, Missouri
34A27B24C26MOK-1fI. dentatus nymphRabbitKoch Farm, Bollinger County, Missouri
35A28B25C28MOK-1gI. dentatus nymphRabbitKoch Farm, Bollinger County, Missouri
36A25B26C27MOK-2-IDaI. dentatus larvaRabbitKoch Farm, Bollinger County, Missouri
37A29B27C27MOK-2-IDbI. dentatus nymphRabbitKoch Farm, Bollinger County, Missouri
26A18B20C21MOK-2-IDcI. dentatus larvaRabbitKoch Farm, Bollinger County, Missouri
38A30B28C29MOK-3aI. dentatus nymphRabbitKoch Farm, Bollinger County, Missouri
39A31B18C30MOK-3bI. dentatus nymphRabbitKoch Farm, Bollinger County, Missouri
40A17B25C16MOK-3cI. dentatus nymphRabbitKoch Farm, Bollinger County, Missouri
40A17B25C16MOK-3dI. dentatus nymphRabbitKoch Farm, Bollinger County, Missouri
41A32B29C31MOJ-1I. dentatus larvaRabbitJenkins Farm, Bollinger County, Missouri
42A32B20C32MOG-1bI. dentatus nymphRabbitGipsy, Wayne County, Missouri
43A33B30C33MON-1I. dentatus nymphRabbitNUT Junction, Cape Giradeau County, Missouri
44A28B18C25MOH-1I. dentatus femaleRabbitHappy Farm, Bollinger County, Missouri
45A32B31C34MOH-2I. dentatus maleRabbitHappy Farm, Bollinger County, Missouri
46A18B32C17BC-1I. dentatus nymphDragMacon, Bibb County, Georgia
47A34B20C35SI-10I. scapularis femaleDragSapelo Island, McIntosh County, Georgia
B. garinii49A36B34C3720047I. ricinusFrance
Borrelia sp.50A37B35C38TXW-1D. variabilis maleCoyoteWebb County, Texas

Whole-cell DNA was extracted using slight modifications of published procedures [15].

RAPD-PCR reactions were performed using the method described previously with minor modifications [14]. RAPD-PCR reactions were performed in volumes of 50 µl containing 10 mM Tris–HCl (pH 8.8 at 25°C), 50 mM KCl, 4 mM MgCl2, 0.1% Triton® X-100, 200 µM dATP, 200 µM dGTP, 200 µM dCTP, 200 µM dTTP, 2 U Taq DNA polymerase in storage buffer A (Promega), 0.4 µM of each primer and 20 ng of extracted DNA using previously published primers 1254, 1283, and AP13 [14]. Reactions were performed in a GeneAmp® PCR System 9700 (PE Biosystems). Briefly, first at 94°C for 2 min and then three cycles of denaturation at 94°C for 5 min, annealing at 36°C for 5 min, and extension at 72°C for 5 min and then 30 cycles of denaturation at 94°C for 1 min, annealing at 36°C for 1 min, and extension at 72°C for 2 min and a final incubation at 72°C for 10 min.

The amplified DNA fragments were separated in 1% agarose gels containing 0.5 µg ml−1 of ethidium bromide in 0.5×Tris–borate–EDTA. For each experimental run, bacteriophage DNA digested with BstEII was included and used as a size marker for the amplified fragments and as a reference for the normalization of different gels. The gels were photographed using an Eagle Eye II System (Stratagene). The PCR-RAPD patterns were measured and analyzed by using EagleSight® software (version 3.2) as recommended by the manufacturer. Each band on the gel was scored as 1 (present) or 0 (absent). The bands with a faint intensity which were not reproducible in duplicate experiments were excluded in the final analysis. Any clear and neat band which was repeatable in duplicate experiments was scored. The scores for three experiments (one experiment performed with each of the three primers) yielded the data matrix. The banding patterns obtained from three primers were combined. The combined patterns were used for the similarity estimation and phylogenetic analysis. The similarity among strains was estimated by neighbor-joining [16]. Isolates with similarity of more than 0.95 were assigned to the same RAPD type. RAPD patterns were converted into strings of 129 characters, representing all possible migration positions of unambiguous RAPD bands. Strings were analyzed by neighbor-joining [16], maximum likelihood [17], parsimony methods [18], and UPGMA (the unweighted pair group method with arithmetic means analysis) [19] using PAUP (Phylogenetic analysis using parsimony) [18].

2.2 Amplification of rrs genes, purification of PCR products, sequence analysis, and phylogenetic analysis

The rrs gene was amplified using previously described primers [20]. PCR reactions were performed in a GeneAmp® PCR System 9700 using conditions described previously [20].

The PCR products were purified using the QIAquick® gel extraction kit (Qiagen) as recommended by the manufacturer. The DNA sequence of the rrs gene was determined using an ABI Prism (model 377). Phylogenetic analysis was performed using PAUP.

2.3 Nucleotide sequence accession numbers and retrieved sequences

The rrs nucleotide sequences of 17 B. burgdorferi s.l. isolates were deposited in the GenBank database, and their accession numbers are B. burgdorferi S.S. [MI-2 (AF467964), SI-1 (AF467958), SM-1 (AF467960), SCI-2 (AF467961)], B. bissettii [MI-6 (AF467966), MI-8 (AF467967), MI-9 (AF467968), SCGT-8a (AF467969), SCGT-10 (AF467970), AI-1 (AF467962), FD-1 (AF467963)], B. andersonii [SI-10 (AF467959), BC-1 (AF467973), MOK-3a (AF467974), MOS-1b (AF467975), MOD-5 (AF467972)], Borrelia sp. [TXW-1 (AF467976)].

To compare the relationship between our strains and the strains in different genospecies or species, and to conduct phylogenetic analyses, rrs gene sequences of strains from the following species/genospecies (with strain names and database accession numbers) were used. They included B. burgdorferi S.S. [B31T (U03396)], B. garinii [20047T (D67018)], B. afzelii [DK1 (X85190)], B. bissettii [DN127T (L40596), 25015 (AJ224138)], B. andersonii [21038T (L46701)], B. valaisiana [VS116T (X98232)], B. lusitaniae [Poti B2T (X98228)], B. japonica [HO14T (L40597)], B. tanuki [Hk501T (D67023)], B. turdi [Ya501T (D67022)], B. sinica [CMN2 (AB022145)], B. hermsii [HS1T (U42292)], B. recurrentis [A5 (AF107357)], B. turicatae [M2007 (U42299)], B. parkeri [6232 (AF307100)], B. lonestari [Texas20 (U23211)], B. coriaceae [Co53T (U42286)], B. miyamotoi [HT31 (D45192)], B. duttonii [Ma (AF107366)], B. hispanica [UESV/246 (U42294)], B. crocidurae [UESV/1096TEN (U42302)], B. persica [UESV/340 (U42297)], Borrelia sp. [Spain (U28502), Florida dog strain FCB-1 (L37837), LB2001 (AY024345)].

3 Results

3.1 RAPD results

Repeated RAPD analyses were performed with reference strains B-31, SH-2-82, DN127, 25015, and 21038 in order to assess the reproducibility of our RAPD fingerprinting procedure. We obtained similar DNA fingerprints of these strains in repeated RAPD analysis. In addition, all strains mentioned above were used as positive controls in each run of RAPD amplification. No differences were seen between the RAPD patterns for B31 at different passages (3 and 8) (data not shown).

We scored 129 polymorphic characters and used them to construct a data matrix. Considerable genetic heterogeneity was noted among the isolates. Based on the RAPD profiles in three individual amplifications with different primers, a total of 44 RAPD types were identified from the 53 southern strains (Table 1). Twenty-nine RAPD types were found among 37 B. andersonii strains, seven RAPD types in eight B. bissettii strains, and seven B. burgdorferi s.s. strains yielded seven RAPD types. TXW-1 formed a separate RAPD type (Fig. 1).

Figure 1

Neighbor-joining phylogenetic tree based on combined gels after three individual amplifications with different primers. The tree was compared with trees produced by maximum likelihood, parsimony, and UPGMA and the four methods produced similar results. Scale bar=calculated distance value. Bootstrap confidence levels above 50% are indicated to the left of each relevant cluster. Numbers in parentheses to the right of each strain indicate different RAPD types and indistinguishable isolates within each genospecies.

Phylogenetic trees were constructed based on the combined gels obtained after three individual amplifications with different primers. A general phylogenetic tree derived from combined data arranged the 53 Borrelia strains into four genetic groups: B. andersonii, B. bissettii, B. burgdorferi s.s., and TXW-1. Greater genetic diversity occurred among B. burgdorferi s.l. strains (six clusters/seven RAPD types/seven strains) compared to B. andersonii strains (13 clusters/29 RAPD types/37 strains) and B. bissettii strains (four clusters/seven RAPD types/eight strains) (Fig. 1).

The phylogenetic trees derived from individual amplifications with the three different primers agreed with each other at the genospecies level, but not at the strain level (data not shown but are available upon request). For example, compared to B. burgdorferi S.S. strains SI-14 and SCI-4 and B. bissettii strains MI-6, and MI-9, FD-1, more B. andersonii strains shared common banding patterns in the gels after amplification with primer 1283. Similar results were obtained after two other amplifications with primers 1254 and AP13. Although the phylogenetic trees based on individual amplifications using different primers reflect the phylogenetic relationships among strains at the genospecies level, the tree derived from combined data provides a better understanding at the strain level (Fig. 1).

3.2 Analysis of rrs sequences

There is a high level of rrs sequence conservation at the species level among those we examined. Of 1337 bp of rrs sequences that were compared among strains in the 11 species in the B. burgdorferi s.l. complex, 80 variations were found. Most variations are replacements except for six deletions in B. japonica, one deletion in B. turdi, three deletions in B. sinica, and one insertion in B. andersonii strains 21038, BC-1, MOS-1b, and MOD-5. Some variable signature positions are distinct in different species in this complex (data not shown but are available upon request). Compared to B. burgdorferi s.l. complex, more mismatches in nucleotide sequences exist among strains of relapsing fever borreliae. We found 98 variations among the sequences in strains of 16 species. There was one deletion in strain Spain, one insertion in strain Spain, B. persica, and strain LB2000, and 93 mutations among the 16 species. Some mutations were different among distinct species (data not shown but are available upon request).

Nucleotide identities were in the range of 96–100% within the B. burgdorferi s.l. complex, 94–99% within relapsing fever borreliae, and 93–99% between the two species complexes (data not shown but are available upon request). The relapsing fever spirochete TXW-1 segregated with the North American relapsing fever borreliae and formed a separate subbranch (Fig. 2). The similarities of nucleotide sequences of rrs among TXW-1 and B. turicatae and B. parkeri are 99%. The DNA sequences of rrs from different genospecies/species in B. burgdorferi and relapsing fever borreliae complex are highly homologous, ranging from 95.3 to 99.6%[20]. However, 0.4–4.7% variation in rrs sequences is enough to infer the taxa of Borrelia and a value of 99% identity is sufficient to differentiate species (data not shown but are available upon request) [2026]. TXW-1 is close to, but separate from B. turicatae, B. parkeri, and the isolate from the Florida dog (Fig. 2).

Figure 2

Phylogenetic tree derived from 16S rDNA nucleotide sequences of southern and reference strains of B. burgdorferi s.l. and relapsing fever borreliae. The neighbor-joining tree was constructed with PAUP software and based on a comparison of 1537-bp nucleotide sequences of near complete genes. The scale bar represents the calculated distance value. Bootstrap confidence levels above 50% are indicated to the left of each relevant cluster. Boldface strain designations indicate the novel sequences determined in this study.

All the strains analyzed here can be divided into two parts: the B. burgdorferi s.l. complex and the relapsing fever borreliae complex (Fig. 2). In the former, there are eight branches. The relapsing fever borreliae consist of six branches. The branching of the relapsing fever borreliae in the phylogenetic tree reflects the geographic distribution and vector diversity of these organisms (Fig. 2). TXW-1 forms a separate subbranch. The genetic distance (indicated by branch length) between TXW-1 and B. turicatae (0.002) is not only greater than the genetic distance between B. parkeri and B. turicatae (0.001), but also greater than the genetic distance between some other members in this complex, such as between B. miyamotoi and B. lonestari (0.001) (data not shown but are available upon request).

4 Discussion

RAPD analysis has been used widely for molecular typing and genetic characterization of various microorganisms [11]. Its reliability and reproducibility have been confirmed by previous study [1214] and our present results. Previous studies indicated that one strain among 136 isolates analyzed has one band change when different clones were used [14], but RAPD fingerprints from different colonies still showed more than 95% similarity to each other and belonged to the same RAPD type. Whether this difference is caused by a mixed population of two strains or by colonial variation within one strain was not clear at that time [14]. We think the banding pattern change was most likely caused by two mixed strains rather than colonial variation. Genomic sequence data indicated that B. burgdorferi s.l. strain B31 contained a linear chromosome of 910 725 bp and at least 17 linear and circular plasmids with a combined size of more than 533 000 bp [27,28]. Some plasmids might be spontaneously lost during in vitro cultivation. However, previous study and our present experiment indicated that the number of passages did not affect the results of RAPD analysis [14]. The reason is that RAPD targeted a few DNA sequences that may or may not be located in the lost plasmids. All isolates used in this study have been cloned and were the same passage. Results obtained from RAPD or AP-PCR agree with the phylogenetic analysis derived from other molecular methods, such as multilocus enzyme electrophoresis, DNA–DNA hybridization, and RFLP of certain genes [2,1214]. Data derived from our RAPD analyses agree with data obtained from pulsed-field gel electrophoresis (PFGE), RFLP and sequence analysis of ospC, rrs, and the rrf-rrl intergenic spacer [9,10,29].

Although RFLP and sequence analysis of the rrf-rrl intergenic spacer are efficient methods for identification of B. burgdorferi s.l. [30], our characterizations using these methods with some of the southern strains have not been totally satisfactory. As noted in our previous report [9], the taxonomic status of several strains, including TXW-1 and MI-8, was uncertain and needed further analysis by other techniques. Strains within genospecies were not differentiated based on RFLP and sequence analysis of the rrf-rrl intergenic spacer. RAPD and sequence analysis of the 16S rRNA gene appear to be the appropriate techniques, since both methods can be used to identify strains not only in the B. burgdorferi s.l. complex, but also in the relapsing fever borreliae complex. Previous data indicated that TXW-1 appeared to be closer to the B. garinii genospecies than to other genospecies [9]. The reason that this similarity incorrectly appeared was that not all genospecies in the B. burgdorferi s.l. complex and all species in the relapsing fever borreliae complex were included in our previous analysis. Thus, TXW-1 appeared incorrectly close to B. garinii. Present studies provide evidence of the novelty of TXW-1 and suggest that it might be a novel strain in the relapsing fever borreliae complex. The taxonomic status of another southern strain, MI-8, was uncertain in our previous report [9], but both RAPD and sequence analysis of rrs indicate that this strain is a member of the B. bissettii genospecies.

Results based on RFLP and sequence analysis of the rrf-rrl intergenic spacer did not differentiate strains within genospecies, especially for B. andersonii [9]. Since RAPD is a strain-specific technique, most southern strains were differentiated very well. These data are useful in analysis of the relationship among genospecies of Borrelia and their tick vector, mammal reservoirs, geographic distribution, and pathogenicity.

Because RAPD analysis randomly samples sequence polymorphisms distributed throughout the genome, the fingerprints are accurate indicators of genetic distances and therefore can be used to establish phylogenetic relationships at interspecific and intraspecific levels. Moreover, because RAPD data take the form of discrete characters, it is easy to add new taxa to phylogenetic distance trees. Compared to results obtained by multilocus enzyme electrophoresis and other molecular methods, the RAPD protocol is simple and the number of polymorphisms almost unlimited.

RAPD types are correlated with geographic locations. For example, B. bissettii strains from Florida and South Carolina were separated into different clusters, and B. burgdorferi S.S. strains from South Carolina and Florida were separated from strains isolated in Georgia. Studies based on RAPD and PFGE analyses indicated that B. burgdorferi S.S. strains from North America separated from strains from Europe [13,14].

Previous studies suggested that RAPD types might be used to infer possible pathogenic potential for B. garinii strains, but not for B. afzelii strains [14]. B. garinii strains from patients with disseminated disease were recognized as different from strains from ticks. However, RAPD patterns of B. afzelii strains from ticks and humans were randomly distributed. In addition, strains from patients with Acrodermatitis Chronica Atrophicans (ACA) were not recognized as different from strains from patients with Erythema Migrans (EM) [14]. These results agree with a study based on PFGE [29]. In our present study using RAPD, southern B. burgdorferi S.S. strains formed several branches (Fig. 1). Southern strain MI-2 formed a separate branch among the B. burgdorferi S.S. strains. This agrees with a recent study based on ospC sequence analysis in which MI-2 most closely aligns with a so-called invasive cluster that possesses many proven human-invasive strains. This suggests that MI-2 might be invasive for humans [10]. Transmission experiments showed that this group of southern spirochetes is able to infect mice and hamsters and that Ixodes scapularis can transmit it (Oliver et al., unpublished data). Further studies are needed to determine if they are pathogenic for humans.

Sequence analysis of rrs has been widely used for species identification, phylogenetic and evolutionary analysis of most bacteria including B. burgdorferi s.l. [18,2731]. At least seven of 11 genospecies were described based on the sequence analysis of 16S rRNA [18,2731]. Different species or genospecies of Borrelia isolates obtained from diverse geographic regions and various biological sources fall into individual clusters in the phylogenetic tree based on 16S rRNA sequence analysis [2026]. Therefore, sequence analysis of the 16S rRNA gene was considered as a common method for identification of most bacteria including B. burgdorferi and represents a reliable method for inferring the taxa of B. burgdorferi s.l. [2026]. Sequence analysis of the 16S rRNA gene provide a useful tool to identify and classify Borrelia, not only B. burgdorferi s.l., but also relapsing fever borreliae. All genospecies/species that we analyzed in the B. burgdorferi s.l. and relapsing fever borreliae complex were differentiated as distinct entities in the phylogenetic tree based on sequence analysis of rrs.

Most relapsing fever borreliae are transmitted by soft-bodied ticks (Ornithodoros or Argas ticks) and one species by the human body louse [3]. In contrast, four recently identified Borrelia, B. miyamotoi [31], B. lonestari [32], an unusual northeastern USA Borrelia [33], and TXW-1 are transmitted by the hard-bodied ticks, Ixodes persulcatus, Amblyomma americanum, I. scapularis, and probably Dermacentor variabilis, respectively. Interestingly, although B. miyamotoi was described from Japan [31], a genetically closely related species was recently reported in the USA [33]. Thus far, B. lonestari and TXW-1 have been found to occur only in the USA. Although human infection by B. lonestari has been reported [32,34], it is unknown whether the other species cause illnesses in humans.

Studies on the relationship between ospC of Lyme disease spirochetes and vmp of relapsing fever borreliae (B. hermsii) suggest that B. burgdorferi and relapsing fever borreliae might have evolved from a common ancestor, but have adapted to different arthropod and vertebrate hosts [3537]. In general, Lyme disease spirochetes have adapted to different species of hard ticks and relapsing fever borreliae have adapted to distinct species of soft ticks, and one Borrelia species to a louse. There appears to be a rather loose association between different genospecies of B. burgdorferi and certain species of Ixodes ticks, except for B. andersonii and I. dentatus which appears to be quite specific. Compared to Lyme disease spirochetes, relapsing fever borreliae have a stricter specificity for their tick vectors. Nevertheless, our results and other reports suggest that relapsing fever borreliae can be transmitted by hard ticks. Further studies on the diversity of Borrelia and their adaptation to ticks under experimental conditions are needed to better understand possible co-speciation among them.

Acknowledgements

We thank Dr. Mark Schembri, Editor of FEMS Microbiology Letters, and three anonymous reviewers for constructive suggestions to improve the manuscript. This research was supported in part by Merit Grant R37 AI-24899 from the National Institutes of Health and Cooperative Agreement U50/CCU410282 from the Centers for Disease Control and Prevention. The opinions expressed are the responsibility of the authors and do not necessarily represent the official views of the NIH or CDC.

References

  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
  26. [26].
  27. [27].
  28. [28].
  29. [29].
  30. [30].
  31. [31].
  32. [32].
  33. [33].
  34. [34].
  35. [35].
  36. [36].
  37. [37].
View Abstract