The Family Borreliaceae (Spirochaetales), a Diverse Group in Two Genera of Tick-Borne Spirochetes of Mammals, Birds, and Reptiles

Abstract

Spirochetes of the family Borreliaceae are, with one exception, tick-borne pathogens of a variety of vertebrates. The family at present comprises two genera: Borrelia (Swellengrebel), which includes the agents of relapsing fever, avian spirochetosis, and bovine borreliosis, and Borreliella (Gupta et al.), which includes the agents of Lyme disease and was formerly known as ‘Borrelia burgdorferi sensulato complex’. The two genera are distinguished not only by their disease associations but also biological features in the tick vector, including tissue location in unfed ticks and transovarial transmission. Borrelia species transmitted by argasid (soft) ticks tend to have more exclusive relationships with their tick vectors than do other Borrelia species and all Borreliella species that have ixodid (hard) ticks as vectors. The division of genera is supported by phylogenomic evidence from whole genomes and by several specific molecular markers. These distinguishing phylogenetic criteria also applied to three new species or isolates of Borrelia that were discovered in ixodid ticks of reptiles, a monotreme, and birds. Although the deep branching of the family from other spirochetes has been a challenge for inferences about evolution of the family, the discovery of related microorganisms in the gut microbiota of other arachnids suggests an ancestral origin for the family as symbionts of ticks and other arachnids.

Epigraphs:

There is still little agreement as to accepted species, and even some generic designations are not accepted by many workers. This is particularly true of the genus Borrelia …’ Gordon Davis, 1948 (Davis 1948)

‘The concept is expressed that Borrelia developed as symbionts of ticks (especially Argasidae) but act as parasites in mammals and birds, which serve as borrelial amplifiers following bites by infected ticks.’ Harry Hoogstraal, 1979 (Hoogstraal 1979)

For the vast majority of time since ancient Greeks began placing animals and plants into different coherent categories, the ‘characters’ for these typologies have exclusively been phenotypes, primarily manifested as discernible differences between organisms in structure, development, behavior, and life history. Bacteria were not isolated in the laboratory until the nineteenth century, but for most of the subsequent time, phenotypes were the only characters used for taxonomy. These included cell morphology under the microscope, colony types, reactions with stains, and differential growth on culture media, among others (Stanier and Van Niel 1941). Eventually other definable and distinguishing phenotypes were added to identification keys for bacteria. These more clearly represented what were genetic determinants but still were removed from the genes themselves. These additional characters of bacteria included ultrastructural features, such as the presence of one or two cell membranes, and biochemical capacities, such as those conferred by specific enzymatic activities, like urease, or the absence or presence of certain pathways, such as oxidative phosphorylation or nitrogen fixation (Buchanan and Gibbons 1974).

In the mid-20th century, when Gordon Davis of the first epigraph was at the National Institutes of Health’s Rocky Mountain Laboratory, investigators likewise relied sets of distinctive phenotypes for systematics. For the arthropod vectors under study, the discriminating traits were inferred from detailed descriptions, aided by dissecting scopes, drawings and photographs, of the anatomy of different developmental stages. These were systematized as morphological keys still in wide use today. Not until 20th century’s end was there application of molecular-based typing measures based directly on the genetic content of the organisms. These began with electrophoretic migrations of enzymes located with colorimetric substrates and with DNA–DNA hybridizations. These low resolution approaches were in time supplanted by the sine qua non of the molecular or phylogenetic approach for systematics: DNA sequences of informative genetic loci, mainly mitochondrial genomes and mitochondrial and chromosomal ribosomal RNA genes, with whole genome sequences in the wings to be the references of record.

In the case of the bacteria transmitted by these arthropods, mid-century microbiologists and entomologists were, if anything, more handicapped in their classification attempts. Most of the known agents of human disease were either obligate intracellular organisms, such as Ricksettia, Orientia, Ehrlichia, and Anaplasma, which only could be examined while they were inside of other cells, or extracellular bacteria, like the spirochetes that caused relapsing fever, that had yet to be cultivated outside of a vertebrate host. Other arthropod-associated bacteria were endosymbionts, entirely confined to the insect or tick hosts for their lives and even less accessible to study. (An exception among vector-borne pathogens was Yersinia pestis, a facultative intracellular pathogen that could be readily cultivated in the laboratory and early on was suitable for the application of microbiologic techniques developed for other proteobacteria, like Escherichia coli.)

What that era’s entomologists and microbiologists could achieve with the vector-borne bacteria, beyond noting disease associations and their locations in their host’s tissues and cells, was application of ecologic features, such as associations of the bacteria with different vectors and reservoirs in nature. This led to investigations in the laboratory of the specificities of those relationships, mainly in terms of the competencies of the candidate arthropod vector and the vertebrate reservoir to serve these roles (reviewed in Davis 1952, Hoogstraal and Aeschlimann 1982, Burgdorfer and Hayes 1989, Schwan and Piesman 2002, Barbour 2004). These studies entailed testing the exclusivity of an association by assessing the ability of other arthropods of same type, e.g., other species of soft-bodied ticks, or different types, e.g., biting flies, to serve as a vector for the microbe. For the vertebrates it meant investigations of the susceptibility of an animal for the microbe and the natural history of infection that ensued. One example of a subject of such studies is Borrelia duttonii, an agent of tick-borne relapsing fever (RF) in sub-Saharan Africa. The soft-tick or argasid Ornithodoros moubata Murray (Ixodoida: Argasidae) was known to be the principal vector of the pathogen for humans. O. moubata was confirmed in experimental studies to be a competent vector for B. duttonii, but another argasid species O. turicata (Duges) of North America was not (Fig. 1). While O. moubata was not competent as a vector of B. turicatae of North America, B. latyschevi of central Asia and Russia, or the avian borreliosis agent B. anserina, it was capable of acquiring and sustaining two other relapsing fever agents: B. crocidurae and B. hispanica. These are two other ‘Old World’ RF agents of the Afrotropical and Palearctic regions, respectively. These experimental results pointed to a greater similarity of B. duttonii to B. crocidurae and B. hispanica than to the other three species included in these studies.

Fig. 1.

Borrelia-tick vector relationships. The names of ticks species are horizontally arrayed, with argasid (soft) ticks on the left and ixodid (hard) ticks on the right. The names of Borrelia species are vertically arrayed. The established associations of a spirochete species and a particular tick as its usual vector in nature are indicated by the diagonal of bordered cells with ‘+’ in them. When there were experiments on the vector competencies of other ticks for a given Borrelia species, the outcomes of the study are indicated by ‘+’ or ‘ –’, where ‘+’ means that the tick species was shown to be competent as a vector. When there was acquisition of a Borrelia species by an atypical tick but with an apparent decline in fitness of the tick, this is indicated by ‘±’. Whether transovarial transmission was determined to occur (+), not occur (−), or has not been reported (?) is indicated in the second column from the left. Borrelia species with valid names are italicized.

Such studies, carried out in numerous laboratories across the world, including Africa, the Middle East, Russia, and Central and South America, yielded an incomparable amount of information about pathogen–vector–reservoir host relationships and interactions. These observations not only advanced understanding of the biology and ecology of these agents, but the individual specificities and collective profiles were of major utility for systematics, as earlier taxonomy reference works for bacteria demonstrate (Murray and Davis 1948). In the case of the agents of RF, it is very difficult to discern a difference between B. hermsii and B. persica, or even B. anserina or B. miyamotoi, which is transmitted by hard ticks, by light microscopy of a stained smear or wet-mount of the blood of an infected person or other animal. They all look the same. With development by Kelly of a culture medium that allowed serial cultivation of RF agents (Kelly 1971), one could discern differences between species in their capacities to grow in this medium or its successors. But once growth was achieved, the appearances of the bacteria of different species as they swam about the broth medium or were fixed on slides were largely indistinguishable.

What could be more easily accomplished after isolation and cultivation of the spirochetal agents of RF was electron microscopy, which revealed ultrastructural characteristics, such as the number of periplasmic flagella they had at each end, or the tightness or laxity of the coils (Holt 1978). These confirmed that the relapsing fever agents were a distinctive and coherent group of species and differentiated from other types of spirochetes, such as Treponema pallidum, the cause of syphilis, or Leptospira interrogans (Hovind-Hougen 1974). But they were at the time of little added value for distinguishing one RF species from another. As discussed below, light microscopy and ultrastructural studies did turn out to be useful for distinguishing the RF agents from what at the time (i.e., 1980s) were the newly discovered agents of Lyme disease, now members of the genus Borreliella.

We are now in a time when genome sequencing for systematics, as well as for many other types of biological research, is standard practice and usually required for acceptance of a formal description of a new species of bacteria. The most critical part of the claim for a new species of bacteria is usually its genome sequence and both how similar it is to other members of the genus and, at the same time, how different it is from other species in the genus. It is uncommon now for such descriptions of new species to include more than a few comments on the vector and reservoir associations of the candidate species in nature or from medical experience. The experimental studies of vector and vertebrate specificities and host ranges for a newly recognized organism are seldom part of these descriptions. Electron microscopy studies of the ultrastructure are also less common in the literature and may not be expected for new descriptions.

As a consequence of this changing emphasis over time of what is most important for systematics-from phenotype to genotype—this selective survey of Borreliaceae deals with heterogeneous information on the various species. Some species of Borrelia, such as B. mazzottii, are described in the literature with enough supporting data about their phenotypes—in terms of vector and vertebrate associations at least—to merit them with their own species designation (Mazzotti 1949, Davis and Mazzotti 1953, Davis 1956b), but there is no isolate of that organism or any DNA sequence for it. At the other extreme are some species of Borreliella, e.g., Bo. finlandensis (Casjens et al. 2011), for which there are complete genome sequences, but comparatively little is known about the range of competent vectors and reservoirs. Between these two extremes is an organism such as B. theileri, about which there is an extensive literature on its ecology, its tick associations, and infections of mammals (Theiler 1904, Smith et al. 1978, Smith et al. 1985, McCoy et al. 2014). There are a handful of DNA sequences for B. theileri in the database but not a genome sequence.

Lengthy descriptions and supporting data about the family Borreliaceae and its two genera have been published to date (Gupta et al. 2013, Adeolu and Gupta 2014, Barbour 2018, Barbour and Schwan 2018, Barbour and Qiu 2019). There also have been reviews of bacteria–tick associations for the family, including monographs by Felsenfeld and Hoogstraal, which predated the discovery of the Lyme disease agents (Felsenfeld 1965, Hoogstraal 1979). The present article instead balances a summary of phenotypic differences between the genera, with a focus on the bacteria–tick associations, and an update on the phylogenetics which includes some newly described species in the family that had not been included in surveys before. We end with a more speculative consideration about the origins of the family.

Phenotypic Traits in the Age of Sequence-Based Systematics

Many of the characteristics that once were considered important for placing a tickborne spirochetal pathogen in a taxonomic group are seldom included in the formal descriptions of new species in the family Borreliaceae by their proposers. While there may be some narrative about the tick and vertebrate associations in nature and possible relationships with human or domestic animals diseases, more in depth studies in the laboratory of these associations, as might have occurred in the past, are often lacking. Ultrastructural studies are also uncommonly included in new descriptions of new species. As a consequence, much of the following summary of biological and phenotypic differences between genera and species pertains to the species that were known before the DNA sequence era.

A difference of medical and epidemiologic importance between the two genera is this: the genus Borreliella includes all the known species that cause Lyme disease, and the genus Borrelia includes the known species that cause relapsing fever (Table 1). There are other species in each genus that are not known to cause any human disease, but these disease associations are stand-out features of the two clades. Lyme disease and relapsing fever in humans have their counterparts in natural and experimental infections in other vertebrate hosts, and these are likewise distinguishing (reviewed in Barbour and Hayes 1986, Barbour 1987, Cadavid and Barbour 1998, Barthold et al. 2010, Crowder et al. 2016). Infections of mammals by Borreliella species tend to be persistent and with higher densities of spirochetes in tissues other than blood (Nakayama and Spielman 1989, Bunikis et al. 2004, Barbour et al. 2009). Borreliella burgdorferi reservoirs may remain infected for their entire lives. In contrast, during relapsing fever in humans, experimental animal infections with relapsing fever species, avian spirochetosis caused by B. anserina, and bovine borreliosis caused by B. theileri, spirochetes in blood are numerous enough for detection by microscopy of thin blood smears to be a common method for laboratory diagnosis of these infections (reviewed in Assous and Wilamowski 2009). Humans and animals become acutely ill, often with signs of sepsis. They may experience several relapses of illness, but once those have resolved, it is difficult to find evidence of residual infection. The Lyme disease agent Bo. mayonii of the north-central United States was reported to reach higher densities in the blood in infected patients than expected for Lyme disease caused by Bo. burgdorferi (Pritt et al. 2016). But even at cell densities of 10⁴–10⁵ per milliliter of blood for Bo. mayonii, this would still be hundred-fold lower than typically observed during relapsing fever in humans. In addition, Bo. mayonii spirochetes were not detected in the blood of experimentally infected mice at a time when spirochetes persisted in the skin and mice remained infectious for ticks (Dolan et al. 2017). This pattern of few spirochetes in the blood and more abundant in other tissues corresponds to the phenotypes of other LD species.

Table 1.

Selected phenotypic characteristics of Borreliaceae genera

Characteristic	Borrelia	Borreliella
Associated disease(s)	Tick-borne relapsing fever and louse-borne relapsing fever; avian spirochetosis; bovine borreliosis; possible disease in reptiles and monotremes	Lyme disease (Lyme borreliosis)
Arthropod vectors	Argasid ticks, prostriate ixodid ticks, metastriate ixodid ticks, and louse (insect)	Prostriate ixodid ticks
Presence in salivary gland in unfed tick	Yes	No
Transovarial transmission in ticks	Most tick-borne species	No
Number of subterminal insertions of flagella at each end	15–20	7–11
Coils of cells	Regularly spaced	Regularly or irregularly spaced
Aggregates of cells in liquid medium	No	Yes

One of the ecologically relevant traits that distinguishes Borrelia and Borreliella spirochetes is the location of the spirochetes in an unfed tick before it embeds and the bloodmeal commences (reviewed in Schwan and Piesman 2002, Piesman and Schwan 2010). Except for the special case of B. recurrentis, which is not transmitted by its louse vector through saliva (Piesman and Schwan 2010), all the other RF group species that have been thus characterized had spirochetes disseminated from the midgut to the salivary glands before the subsequent molt and the next bloodmeal. These include the Borrelia species in rapid-feeding argasid ticks: B. duttonii in O. moubata (Burgdorfer 1951), B. anserina in Argas spp. (Diab and Soliman 1977), B. crocidurae in Ornithodoros spp. (Gaber et al. 1984), B. hermsii in O. hermsi Wheeler (Schwan and Hinnebusch 1998), and B. turicatae in O. turicata (Lopez et al. 2013, Krishnavajhala et al. 2017). But this feature also applies to Borrelia spp. in the slow-feeding ixodid or hard ticks: B. theileri in Rhipicephalus (Boophilus) microplus (Canestrini) (Ixodida: Rhipicephalinae) (Smith et al. 1978), B. turcica in Hyalomma spp. (Takano et al. 2010), and B. miyamotoi in Ixodes scapularis Say (Ixodida: Ixodinae) (Linda Bockenstedt, personal communication).

This behavior of Borrelia species contrasts with the location of Borreliella species in unfed vector ticks, all of which are species of the genus Ixodes. A workshop consensus statement in 1991 stated that ‘… the Lyme disease spirochete in its Ixodes vectors remains in the midgut, where it aggregates near the microvillar brush border and in the intercellular spaces of the gut epithelium’ (Burgdorfer et al. 1991). Migration from the midgut through the hemolymph to the salivary glands begins after the commencement of the bloodmeal. Subsequent studies confirmed this (reviewed in (Schwan and Piesman 2002, Piesman and Schwan 2010). Previous reports of dissemination of a Lyme disease agent in unfed ticks were before the discovery of B. miyamotoi and represented misidentification of B. miyamotoi as Bo. burgdorferi by nonspecific stains or polyclonal antisera that was cross-reactive among members of the family Borreliaceae. According to Piesman in his report on experimental transmissions: while Bo. burgdorferi spirochetes were ‘occasionally detected in salivary glands early during tick feeding, sufficient numbers of infectious spirochetes were not present within the salivary glands to cause infection in experimental hosts until at least 60 h after tick attachment’ (Piesman 1995).

A second trait of importance for the ecology of these infectious agents is transovarial transmission of the bacterium (reviewed in Burgdorfer and Varma 1967, Barbour 2004, Rollend et al. 2013; Table 1). This alternate means of transmission to another generation of ticks serves to maintain a population in the absence of an infected vertebrate host. While most Borrelia species feature transovarial transmission in their tick vectors (Fig. 1), this phenomenon has not been documented in any member of Borreliella in which this has been investigated. Transovarial transmission of Borrelia species occurs in hard ticks as well as soft ticks. The former include B. theileri in R. microplus (Smith et al. 1978), B. lonestari in Amblyomma americanum (L.) (Ixodida: Amblyomminae) (Killmaster et al. 2014), and B. miyamotoi in I. scapularis (Scoles et al. 2001, Han et al. 2019). Older reports of transovarial transmission of Borreliella species in Ixodes spp. ticks are also attributable to misinterpretation of B. miyamotoi with nonspecific stains and antisera (Rollend et al. 2013).

As noted, ultrastructural studies are seldom included in the descriptions of new species in the family Borreliaceae. However, when this was more routinely performed, the data from those studies support the inclusion of flagella numbers as one of the informative phenotypes for classification. In their 1984 paper Schmid et al. noted that ‘the Lyme disease spirochetes differ from Borrelia spp. morphologically’ (Schmid et al. 1984). Johnson and Hyde and Hovind-Hougen in the same year distinguished between the new spirochetes and known Borrelia spp. in the number of insertion points for flagella they had at their ends (Hovind-Hougen 1984, Johnson et al. 1984). The original observation of comparatively fewer flagella in ‘Lyme disease spirochetes’ than in other Borrelia spp. has held up. The most commonly reported values for flagellar insertion point at one end of cells of Borelliella is 7–8, within a range of 4–12, for the following species: Bo. burgdorferi (Hovind-Hougen 1974, 1995; Barbour and Hayes 1986; Hovind-Hougen et al. 1986; Kudryashev et al. 2009), Bo. afzelii and Bo. garinii (Hovind-Hougen et al. 1986, Hovind-Hougen 1995, Kudryashev et al. 2009), Bo. sinica (Masuzawa et al. 2001), and Bo. japonica (Yano et al. 1997).

With the exception of B. anserina, which is the agent of avian spirochetosis and was reported to have 7–8 flagella at an end (Hovind-Hougen 1995), all other examples of Borrelia that have been adequately examined by ultrastructural criteria have had 15–30 flagellar insertion points at one end. These include B. recurrentis (Hovind-Hougen 1974, Hovind-Hougen 1995), B. hermsii (Barbour et al. 1982, Guyard et al. 2013), a Palearctic species called B. microti (Hovind-Hougen 1995, Naddaf et al. 2012), and B. persica (Karimi et al. 1978). B. anserina is distinguished from these other Borrelia species by having the smallest genome of the group and evidence of lost or degraded genes (Elbir et al. 2017).

A plausible consequence of the flagella number is the regularity and amplitude of the coils of the spirochete (Table 1). Mutants of Bo. burgdorferi which do not make flagella are straight without any coils (Sadziene et al. 1991). Borrelia species, with their greater number of flagella, have more regular amplitude and spacing of the coils than the Borreliella species which have been examined in this regard. The comparative laxness of Bo. burdorferi’s shape and its more leisurely motions in medium may account for the greater tendency of this and other Borreliella species to become entangled and form aggregates during in vitro cultivation (Table 1).

Specificities of Bacterium–Tick and Bacterium–Vertebrate Associations

Pathogen–vector associations tend to be more restricted in range for Borrelia spp. transmitted by soft ticks than for those Borreliaceae species of both genera that are transmitted by hard ticks. Ornithodoros and other argasid ticks commonly are nidicolous and thus limited in the variety of potential vertebrate hosts they would encounter. There may be other ectoparasites, such as lice and fleas, in a nest and burrow, but not likely another species of ticks. This is generally not the case for the questing Ixodes ticks that transmit Borreliella species and B. miyamotoi where they may encounter many types of vertebrates in an environment, including birds and reptiles as well as mammals as suitable sources of a bloodmeal. These vertebrates may have other types of questing ticks that parasitize them, and these other ticks can serve as bridge vectors for maintenance of the spirochete population. Bo. burgdorferi is an example of this lower stringency of its relationships with ticks within genus Ixodes. Although Ixodes cookei Packard and I. holocyclus Neumann were reported as poorly competent vectors of Bo. burgdorferi (Piesman and Stone 1991, Barker et al. 1993), the majority of Ixodes spp. that were tested in the laboratory demonstrated some degree of competency: Ixodes hexagonus Leach (Gern et al. 1991), Ixodes pacificus Cooley & Kohls (Lane et al. 1994), Ixodes spinipalpis Hadwen & Nuttall (Dolan et al. 1997), Ixodes jellisoni Cooley & Kohls (Lane et al. 1999), Ixodes angustus Neumann (Peavey et al. 2000), Ixodes muris Bishop & Smith (Dolan et al. 2000), and Ixodes sinensis Teng (Sun et al. 2003).

Figure 1 summarizes the results of several studies of the competencies of various ticks for Borrelia species reviewed in Barbour (2004). The great majority of these studies were done decades ago when, as has been discussed, the specificity of these relationships was important for distinguishing between species. All of the species for which there is experimental data on the competencies of other ticks are transmitted by soft ticks. The hard tick-transmitted Borrelia spp. sometimes were reported present in other hard ticks in field collections but these rare to infrequent associations were not confirmed by experiments in the laboratory.

The only argasid tick that was reported to be able to experimentally transmit a Borrelia sp. other than its usual pair mate was O. moubata, the usual vector of B. duttonii in sub-Saharan Africa. This tick was able to transmit B. crocidurae and B. hispanica as well as B. duttonii to laboratory animals. As discussed below and shown in Fig. 2, these three species are closely related phylogenetically and could by some definitions based on sequence identity be considered different strains of one species. The two other species that O. moubata transmitted in experiments were B. hermsii and B. parkeri, which are genetically distant from B. duttonii and the other two ‘Old World’ species. However, Davis also noted that the O. moubata ticks infected with B. hermsii or B. parkeri were smaller than uninfected ticks and demonstrated comparatively low oviposition (Davis 1942). Therefore, infection with the nonconcordant species may have lowered the fitness of the ticks and resulted in the assignment of ‘±’ to those relationships (Fig. 1). Gordon Davis did not observe local specificity within species of ticks and Borrelia when ticks and organisms of the same species from different locations in the western United States were paired in different combinations in transmission studies (Davis 1956a). But more recently Krishnavajhala et al. (2018) noted some differences between two isolated populations of O. turicata in their acquisition of B. turicatae from infected mice.

Fig. 2.

A maximum-likelihood phylogenetic tree for genome-sequenced Borreliaceae species based on concatenated sequences of 795 core proteins that are commonly shared by Borreliaceae species. The trees were constructed as described in earlier work (Gupta 2019) and rooted using genome sequences from Spirochaeta africana and Sphaerochaeta coccoides (not shown in the figure). The strains of the species and the accession numbers for the corresponding genome sequences are given in Table 2. All nodes in this tree are supported by 100% bootstrap scores. The species that are transmitted by soft (argasid) or hard (ixodid) ticks are indicated. The names of Borreliella species that are documented to cause Lyme disease are underlined. The scale bar indicates the number of substitutions per amino acid position.

The genetic trait or traits of the soft ticks that conferred a degree specificity to the tick’s competence as a vector appeared to be dominant. The F₁ generation of the mating O. turicata with O. parkeri ticks was able to transmit both B. turicatae and B. parkeri (Davis 1942). The putative determinant of specificity may be expressed in the mouth parts or the midgut of the tick. When B. duttonii after serial passage in mice had lost the capacity to infect O. moubata when they fed a bacteremic mouse, the ticks were rendered infected when the mouse-passaged B. duttonii was injected directly into the hemocoel, thereby, bypassing the mouth parts and gut (Varma 1956).

A restricted vertebrate host range was another character that was less frequently used to classify a relapsing fever Borrelia sp. (reviewed in Felsenfeld 1971, Barbour and Hayes 1986, Barbour and Schwan 2018). Examples of the limitation of Borrelia species to a single type of vertebrate host are B. anserina and birds and of B. recurrentis and humans. Although the majority of relapsing fever agents, with the exception of the louse-borne B. recurrentis, could infect the house mouse Mus musculus and rats, there were differing abilities to infect guinea pigs and rabbits. Similarly, in the genus Borreliella the different capacities to infect birds or rodents is an important distinguishing feature between Bo. afzelii and Bo. garinii (Kurtenbach et al. 2002), which often are sympatric. In fact, the atypical ability of a Bo. garinii strain to readily infect mice was one of the justifications for splitting it off as a new species Bo. bavariensis from other B. garinii strains (Fig. 2; Margos et al. 2009).

Phylogenomics of the Family Borreliaceae

While we stress that disease associations and phenotypic traits like transovarial transmission retain their empirical as well as practical relevance for taxonomic considerations and should not be abandoned, phylogenetics based on DNA sequence, now routinely of the whole genomes, complements the phenotypic considerations and lead to new insights about the evolution of this group of microorganisms. Indeed, the case for recognizing ‘Borrelia burgdorferi sensulato complex’ as a new genus, namely Borreliella, was robustly supported by the phylogenetic case (Adeolu and Gupta 2014). Although several clinicians and microbiologists in the Lyme disease field argued against it (Margos et al. 2018), the retention of ‘Borrelia’ for the genus comprising the relapsing fever, avian spirochetosis, and bovine borreliosis agents was required by the Bacteriological Code for this situation (Lapage et al. 1992). A genus name must be retained for the group of organisms, which contains the type species of the genus. As the type species of the genus Borrelia is B. anserina (Bergey et al. 1925), and as this species is in the relapsing fever clade (Fig. 2), the genus name Borrelia must stay with the relapsing fever clade of species (Wang and Schwartz 2011, Barbour 2018).

The phylogenetic distinctiveness of the Lyme disease group of Borrelia spp. and the relapsing fever group had been noted previously on the basis of individual gene sequences (reviewed in Barbour 2001, 2014). Continued and wide use of the term ‘B. burgdorferi sensulato complex’ or some variation of this (e.g., ‘Lyme disease group’) was de facto recognition of the group’s uniqueness. But that term, as well as the corollary term ‘Borrelia burgdorferi sensustricto’ for the first isolated species of Lyme disease agent, continued to be used, even as the ‘Lyme disease group’ much expanded with newly identified species. A cause of confusion for nonspecialists in the medical and public health fields, as well as patient support and lay advocacy groups, is that the majority of new species expanding the ‘Borrelia burgdorferi sensulato complex’ were not in fact causes of Lyme disease.

The terms ‘borrelia’ for the singular and either ‘borrelias’ or ‘borreliae’ for the plural still apply as common names for bacterial cells of both genera. This follows the precedence of continued use of the term ‘lactobacilli’ as a name for bacterial cells of the various newly emended genera, formerly all within genus Lactobacillus, but still under the family Lactobacillaceae (Zheng et al. 2020). Likewise, ‘borreliosis’ still serves as a name for infections by either genus of the family Borreliaceae, as in ‘Lyme borreliosis’ caused by Bo. burgdorferi or ‘bovine borreliosis’ caused by B. theileri.

Preceding this description of a new genus was a reevaluation of the systematics of these tick-associated spirochetes at the taxonomic level of family (Gupta et al. 2013). The relapsing fever and Lyme disease species had been a genus within the family Spirochaetaceae (Wang and Schwartz 2011), which was a very diverse clade that also included the agent of syphilis Treponema pallidum, symbionts of termite guts, and various free-living spirochetes in a variety of environment (Paster and Dewhirst 2000). The family Borreliaceae is now sister taxon to Spirochaetaceae and as such remains within the order Spirochaetales, the class Spirochaetia, and phylum Spirochaetota (or Spirochaetes) (Barbour 2018).

Borreliaceae was originally described on the basis of a long-branch that separated this group of species from all other Spirochaetales in phylogenetic trees, and identification of large numbers of molecular signatures that were uniquely found in different species from this family (Gupta et al. 2013). The distinguishing molecular markers were of two kinds 1) insertions and deletions within conserved regions (referred to as CSIs) of housekeeping and other important proteins that were specific for the Borreliaceae orthologues and 2) conserved signature genes/proteins (referred to as CSPs), whose orthologues in the order are only present in different Borreliaceae species but not in any other bacteria.

In addition to the accumulating evidence from sequence distances revealed by whole genome sequences (Fig. 2), analysis of CSIs and CSPs also strongly supported the formal division of the relapsing fever and Lyme disease clades of the family Borreliaceae; over one hundred unique and distinctive CSIs and CSPs were identified (Adeolu and Gupta 2014). Pairwise comparisons of the whole genome average nucleotide identity (ANI) and average amino acid identity (AAI) values between different Borreliaceae species also indicated that the species from the RF and LD clades are distinct enough to merit assignment to different genera on a phylogenomic basis as well as the aforementioned biological characteristics (Adeolu and Gupta 2014; Fig. 3).

Fig. 3.

An average AAI matrix based on pair-wise comparisons of the amino acid sequences for 795 core proteins of the family Borreliaceae. Individual cells in the matrix are shaded in a continuous gradient ranging from white at the lowest AAI values (0.67) to dark gray (print) or dark green (online) at the highest AAI values (1.0).

Since the proposal for division of the family Borreliaceae into two genera was made (Adeolu and Gupta 2014), sequence information for a number of new species has been added to public databases. These new sequences represent both clades of the family and also include some species that appear to branch distinctly from the RF and the LD clades of species in phylogenetic trees. These newly described species include Borrelia turcica found in reptiles and hard ticks associated with reptiles (Güner et al. 2004, Takano et al. 2010, Margos et al. 2018, Hepner et al. 2020), Candidatus Borrelia tachyglossi isolated from a tick monotreme echidna (Tachyglossus aculeatus) of Australia (Loh et al. 2016, Gofton et al. 2018), and more recently Borrelia sp. A-FGy-1 (designated as Candidatus Borrelia mahuryensis) isolated from Neotropical ticks of passerine birds (Binetruy et al. 2020). The identification of these novel Borrelia species in sources that had previously drawn little attention raises question whether these species are a part of the genus Borrelia or Borreliella. Alternatively, do these three newly described spirochetes form a third group or clade within Borreliaceae, as suggested by Binetruy et al. (2020)?

The maximum-likelihood phylogenomic tree of Fig. 2 includes all available genome sequences of Borreliaceae species as of this writing (December 2020). This tree is based on concatenated sequences of 795 proteins that are shared by the represented species and is rooted using two species from the family Spirochaetaceae: Spirochaeta africana and Sphaerochaeta coccoides (Table 2). In both this tree and the phylogenetic tree of Binetruy et al. (2020), which was based on concatenated sequences for 590 proteins, the known Borreliaceae species form two main clades corresponding to the Borrelia and Borreliella genera. Whereas the species within the Borreliella clade show limited genetic diversity, much greater genetic diversity is observed within the clade corresponding to the Borrelia genus. This latter clade include species that represent the agents of a relapsing fever, avian spirochetosis, and bovine borreliosis, as well as the recently described species/isolates B. turcica, Candidatus B. tachyglossi, and Candidatus B. mahuryensis, which form deeper branching lineages within this clade.

Table 2.

Species, strains, and accession numbers for genomes used as references

Genus	Species	Strain	Accession	Genome
Borrelia	anserina	Es	GCF_001936255.1	Complete
Borrelia	coriaceae	Co53	GCF_000518125.1	Complete
Borrelia	crocidurae	03-02	GCF_000825665.2	Complete
Borrelia	duttonii	Ly	GCF_000019685.1	Complete
Borrelia	hermsii	CC1	GCF_000956315.1	Complete
Borrelia	hispanica	CRI	GCF_000500065.1	Complete
Borrelia	mahuryensis	A-FGy1	GCA_014084025.1	Complete
Borrelia	miyamotoi	Yekat-1	GCF_002741785.1	Complete
Borrelia	parkeri	HR1	GCF_000512145.1	Complete
Borrelia	persica	No12	GCF_000500045.1	Complete
Borrelia	recurrentis	A1	GCF_000019705.1	Complete
Borrelia	tachyglossi	BC-F10-1268	GCA_003076595.1	Assembly
Borrelia	turcica	IST7	GCF_003606285.1	Complete
Borrelia	turicatae	91E135	GCF_000012085.2	Complete
Borreliella	afzelii	PKo	GCF_000165595.2	Complete
Borreliella	bavariensis	PBN	GCF_014525745.1	Complete
Borreliella	bissettii	DN127	GCF_000222305.1	Complete
Borreliella	burgdorferi	B31	GCF_000008685.2	Complete
Borreliella	californiensis	CA446	GCF_014205885.1	Complete
Borreliella	chilensis	VA1	GCA_000808095.1	Assembly
Borreliella	finlandensis	SV1	GCF_000181875.2	Complete
Borreliella	garinii	BgVir	GCF_000239475.1	Complete
Borreliella	japonica	HO14	GCF_900099615.1	Assembly
Borreliella	lanei	CA28	GCF_014205895.1	Complete
Borreliella	maritima	CA690	GCF_008931845.1	Complete
Borreliella	mayonii	MN14-1539	GCF_001936295.1	Complete
Borreliella	spielmanii	A14S	GCF_000181895.2	Complete
Borreliella	turdi	TPT2017	GCF_008828755.2	Complete
Borreliella	valaisiana	VS116	GCF_000170955.2	Complete
Borreliella	yangtzensis	Okinawa-CW62	GCF_014201775.1	Complete
Sphaerochaeta	coccoides	SPN1	GCA_000208385.1	Complete
Spirochaeta	africana	Z-7692	GCA_000242595.3	Complete

In the tree shown in Fig. 2, as well as that published by Binetruy et al. (2020), B. turcica, Candidatus B. tachyglossi, and Candidatus B. mahuryensis group together. However, the branch separating this clade from the other species in genus Borrelia is very short, which casts doubt about the reliability of a designation of a third clade on the basis of the evidence in hand. In Supp Fig. 1 (online only), we show a phylogenetic tree based on concatenated sequences of DNA gyrase A (GyrA) and DNA gyrase B (GyrB) proteins. In this tree, while B. turcica and Candidatus B. mahuryensis group together, Candidatus B. tachyglossi branches separately and more deeply than all of the species within the Borrelia clade.

The division of the Borreliaceae species into the two main clades corresponding to the genera Borrelia and Borreliella is also supported by other metrics based on whole genome sequences. These include pairwise comparison of ANI and AAI between different genomes and comparison of percentage of conserved proteins (POCP) between different genomes (Adeolu and Gupta 2014, Barbour et al. 2017, Gupta 2019). Figure 3 is an updated AAI matrix based on genome sequences of all available Borreliaceae species. The species from the Borrelia, including the newly added species, and Borreliella genera show much higher sequence similarity to the other species from the same genus than to the species from the other genus and these values are nonoverlapping, thereby further documenting the genetic distinctness between these two groups of species. However, based on the AAI matrix, no clear distinction can be made between the B. turcica, Candidatus B. tachyglossi, and Candidatus B. mahuryensis and other members of the genus Borrelia.

Apart from their phylogenomic branching pattern and the whole genome sequence similarity studies, the strongest evidence showing that the species from these two genera are distinct from each other is provided by large numbers of highly specific molecular markers consisting of CSIs and CSPs, which are uniquely and exclusively shared characteristic of the species from these two genera (Adeolu and Gupta 2014, Barbour et al. 2017, Gupta 2019). The study by Adeolu and Gupta that first proposed the division of Borreliaceae into two genera was based on sequence information for 17 species from this family available at the time (Adeolu and Gupta 2014). There were eight CSIs that were identified as specific for Borrelia and seven CSIs were identified as specific for Borreliella. Since then sequence information has become available for 13 additional members of this family.

A test of the utility of these molecular markers for the categorization or distinguishing of these genera (i.e., the ‘rules’) is whether they retain their utility/specificity in correctly classifying the new organisms as they become available into the described (or novel) categories. We have examined the species distribution (i.e., presence/absence) of different described molecular markers in the newly described species. In the chemotaxis response regulator protein CheY, two different CSIs were identified that distinguished Borreliaceae from other spirochete families (Gupta 2013, Adeolu 2014). Both these CSIs were shown to be present in all 13 newly sequenced species (Supp Fig. 2 [online only]).

The results for one CSI each for the Borreliella and Borrelia genera are presented in Supp Fig. 3 (online only). The CSI specific for Borreliella genus, a glutamic acid insertion in the glucose-6-phosphate isomerase protein, is present in all nine subsequently described species that branch within the Lyme disease group clade (panel A). The other CSIs and CSPs specific for the Borreliella genus are also present in all the newly described species from this clade (Gupta 2019). Likewise, all of the CSIs and CSPs for the Borrelia genus are present in the newly-added B. coriaceae sequence (panel B), but they are lacking in species that are part of the Lyme disease clade (Gupta 2019).

With regard to the more deeply branching species or isolates of the relapsing fever clade, four of the eight previously identified CSIs that were specific for the genus Borrelia are also commonly shared by B. turcica, Candidatus B. tachyglossi, and Candidatus B. mahuryensis, demonstrating their unique features but also maintaining a position within a more broadly inclusive genus of Borrelia (panel B of Supp Fig. 3 [online only]; Gupta 2019). One of these is a six amino acid insertion in a tetratricopeptide repeat domain protein (BDU_327 in B. duttonii Ly and BB_0326 of Bo. burgdorferi B31) of all Borrelia species but absent in all Borreliella species (panel B of Supp Fig. 3 [online only]). The remaining four CSIs are specific for only the relapsing fever group of species, but they are absent in B. turcica, Candidatus B. tachyglossi, and Candidatus B. mahuryensis, confirming their deeper branching position within the genus Borrelia.

In summary, the recent addition to the family Borreliaceae of three species that were isolated from diverse sources further demonstrates the breadth of environments and variety of life cycles represented by the genus Borrelia. The discoveries of new species often pose a challenge to prevailing taxonomies. This was the case for the family Borreliaceae and its two extant genera. While much remains to be learned about these newly described tickborne spirochetes, especially their morphology, vector biology, and pathogenicity for vertebrates, the phylogenetic evidence supports retention of these reptile-, monotreme-, and bird-associated spirochetes in the genus Borrelia.

Evolution of Borreliaceae

The trees of Fig. 2 and Supp Fig. 1 (online only) have roots outside of Borreliaceae, but the branches to the other taxa are very deep and thus of limited utility for inferences about evolution of this family of arthropod-borne spirochetes. One person who thought and wrote about this question was one of those pioneering medical acarologists of the Introduction: Harry Hoogstral (second epigraph). Hoogstraal proposed that Borrelia species may have originally been symbionts of ticks (Hoogstraal 1979, Hoogstraal and Aeschlimann 1982). One justification for Hoogstral’s proposition was the recognized phenomenon of transovarial transmission of Borrelia species in ticks, usually soft ticks but at least one hard tick known at the time of his writing: the bovine borreliosis agent B. theileri in a Rhipicephalus (Boophilus) species. For some species of Borrelia vertical transmission occurs at high enough efficiency that a population of the spirochete might be maintained through several generations of ticks without the need for passing through a vertebrate host (reviewed in Barbour and Hayes 1986). It follows that there may be spirochetes related to Borreliaceae that are true symbionts of other types of arachnids.

But until the recent explosion of novel DNA sequences from the studies of the microbiota of various organisms, there was not much in the sequence database to compare Borrelia species sequences with. On the basis of a 16S ribosomal RNA gene sequence, the nearest relative to Borrelia was for many years considered to be a Cristaspira species, a symbiotic spirochete of bivalve mollusks, such as oysters, and recognized since 1910 (Paster et al. 1996). But recent searches of the DNA sequence repositories, now populated with immense amounts data from the sequencing of gut and other microbiota, revealed evidence of two closer relatives to Borreliaceae in the gut metagenomes of two arachnids (Fig. 4). These were a scorpion (Centruroides limpidus) in Mexico and a spider (Stegodyphus dumicola) in South Africa. There probably are other putative close relatives of Borreliaceae to be discovered among arachnids, including mites and other types of ticks. If some of these ‘missing links’ can be more broadly sequenced, they may provide for more confident inferences about the evolution of Borreliaceae and its ancestors shared with other lineages. Harry Hoogstraal passed away before the era of PCR, inexpensive DNA sequencing, and bioinformatics, but he doubtless would be pleased to know that these technologies are providing for tests of his conjecture.

Fig. 4.

Observed distance phylogram of 1,255 aligned nucleotides of 16S ribosomal RNA genes of selected species of the family Borreliaceae, a Cristispira species of an oyster, uncultured organisms of the gut microbiomes of a scorpion (Centruroides limpidus) and a spider (Stegodyphus dumicola), and as outgroup two representatives of the family Spirochaetaceae: Spirochaeta aurantia and Treponema pallidum. The GenBank accession numbers for each representative of the species are shown. Nodes with bootstrap support 75% or more from 1,000 replicates are indicated. The distance scale bar is the proportion of ungapped sites at which aligned sequences differed.