The Role of Borrelia burgdorferi Outer Surface Proteins

Abstract

Human pathogenic spirochetes causing Lyme disease belong to the Borrelia burgdorferi sensu lato complex. B. burgdorferi organisms are extracellular pathogens transmitted to humans through the bite of Ixodes spp. ticks. These spirochetes are unique in that they can cause chronic infection and persist in the infected human, even though a robust humoral and cellular immune response is produced by the infected host. How this extracellular pathogen is able to evade the host immune response for such long periods of time is currently unclear. To gain a better understanding of how this organism persists in the infected human, many laboratories have focused on identifying and characterizing outer surface proteins of B. burgdorferi. Since the interface between B. burgdorferi and its human host is its outer surface, proteins localized to the outer membrane must play an important role in dissemination, virulence, tissue tropism, and, immune evasion. Over the last two decades numerous outer surface proteins from B. burgdorferi have been identified and more recent studies have begun to elucidate the functional role(s) of many borrelial outer surface proteins. This review summarizes the outer surface proteins identified in B. burgdorferi to date and provides detailed insight into the functions of many of these proteins as they relate to the unique parasitic strategy of this spirochetal pathogen.

Introduction

Lyme disease, or Lyme borreliosis, is an arthropod-borne infection caused by the pathogenic spirochete Borrelia burgdorferi (Benach et al., 1983;Steere et al., 1983). Since its discovery in 1975, during an epidemic of oligoarthritis in children and adults (Steere et al., 1977b), Lyme disease has become recognized as the most prevalent arthropod-borne infection in the United States (Centers for Disease Control, 1996). Lyme disease is typically transmitted to humans by the bite of an infected Ixodes spp. tick and the earliest manifestations include a skin rash, termed erythema migrans, with concomitant flu-like symptoms (Steere et al., 1977a). Infected individuals that do not receive antibiotic therapy are at risk for developing chronic forms of the disease which can result in various disorders of the heart, nervous system, and joints. Although this disease is endemic to the East Coast, Upper Midwest, and Pacific coast of the United States, Lyme disease is also widespread throughout many parts of Europe (Barbour and Fish, 1993;Lovrich et al., 1994). The recent increase in the number of Lyme disease cases being reported from various areas of the United States and Europe, (Barbour et al., 1996;Moody et al., 1998), underscores the importance of generating a new and efficacious Lyme disease vaccine. In this regard, the outer surface lipoprotein A (OspA)-based vaccine for Lyme disease, which was approved for human vaccination for several years, was taken off the market almost a decade ago and is no longer in use. Therefore, the identification of new outer surface proteins that could be used as a second-generation vaccine is now not only warranted for basic scientific reasons, but also is important for overall public health.

Outer surface proteins of B. burgdorferi

Antibodies directed against outer surface proteins (e.g., OspA) have been shown to protect animals and humans from infection with B. burgdorferi (Fikrig et al., 1991). Therefore, a major emphasis in B. burgdorferi research has been to develop a new vaccine that could be used as a safe and effective second-generation preventative against Lyme disease. Since B. burgdorferi is an extracellular pathogen, and humoral immunity has been shown to be protective against this organism, vaccine studies have revolved around identifying borrelial antigens that are (i) surface-exposed, (ii) conserved among different strains and genospecies of Borrelia spirochetes, and (iii) produced during tick transmission and mammalian infection. Any outer surface protein that fulfills these three basic requirements is considered an excellent candidate for vaccine studies. Since the surface of B. burgdorferi is the interface between the host and pathogen during infection, outer membrane proteins also have been implicated as important virulence factors.

As a first step in identifying borrelial proteins that are surface-exposed, many laboratories performed microarray analyses to examine the global response of gene expression in B. burgdorferi after exposure to either temperature-shift or cultivation within a mammalian host environment (Revel et al., 2002;Ojaimi et al., 2003;Brooks et al., 2003;Tokarz et al., 2004). The underlying assumption in these studies, which has been supported by empirical data, is that genes upregulated by temperature will correspond to genes upregulated during tick feeding and transmission to the mammalian host, while genes upregulated during cultivation in a mammalian host correspond to genes upregulated during mammalian infection. Using these two different environmental stimuli, numerous genes that are upregulated during tick feeding and/or mammalian infection were identified. Among the genes observed to be upregulated by temperature and/or mammalian specific signals, over 50 have been shown to encode known or putative leader peptides, indicating that they may encode outer surface proteins (Revel et al., 2002;Ojaimi et al., 2003;Brooks et al., 2003;Tokarz et al., 2004). Further, many of the genes identified were observed to encode hypothetical outer membrane proteins that had not previously been characterized. Therefore, a major goal in the Lyme disease field in recent years has been to further characterize surface-exposed proteins by (i) determining their cellular location throughout the enzootic cycle of B. burgdorferi, (ii) examining their overall conservation among different strains and genospecies of B. burgdorferi, and (iii) assessing their ability to protect mice and non-human primates from experimental Lyme disease. The combined studies have led to the identification of several candidate vaccine molecules and to the identification of many virulence determinants.

Differential expression of B. burgdorferi outer surface proteins

The enzootic life cycle of B. burgdorferi is complex and typically involves horizontal transmission between ticks of the genus Ixodes and wild rodents (Lane et al., 1991). Interestingly, our laboratory as well as others have reported that during nymphal tick feeding profound changes occur in the antigenic profile of B. burgdorferi as it migrates from the tick midgut and salivary glands into mammalian tissue (Schwan et al., 1995;de Silva et al., 1996;Hefty et al., 2001;Hefty et al., 2002b). The reciprocal expression of outer surface protein (Osp) A (downregulated) and OspC (upregulated) that occurs during tick feeding was first reported by Schwan and co-workers in 1995 (Schwan et al., 1995). Subsequent to this seminal report by Schwan and colleagues (Schwan et al., 1995), many laboratories have reported on the identification of several differentially expressed B. burgdorferi antigens, some of which are upregulated by an increase in temperature (Hefty et al., 2001), while others appear to be expressed exclusively during the mammalian phase of infection (Akins et al., 1995;Suk et al., 1995;Wallich et al., 1995;Fikrig et al., 1999;Champion et al., 1994;Hefty et al., 2002b). Although there are exceptions (Aron et al., 1996), almost all differentially expressed B. burgdorferi antigens identified to date are plasmid-encoded (Ojaimi et al., 2003;Brooks et al., 2003). This has led investigators to speculate that these extrachromosomal plasmid elements are essential for both B. burgdorferi virulence and maintenance of the borrelial enzootic cycle. This notion is further supported by the finding that changes in plasmid content correlate with loss of B. burgdorferi infectivity (Purser and Norris, 2000;McDowell et al., 2001;Labandeira-Rey and Skare, 2001).

Prior studies have now shown that many of the borrelial surface antigens are lipid-modified proteins (i.e., lipoproteins). Interestingly, Cox and coworkers noted that several surface-exposed lipoproteins (OspA, OspB, and OspC) are not found exclusively on the surface of the organism. In fact, these lipoproteins can be detected in the periplasm of the organism as well (Cox et al., 1996). Lipoproteins are not only differentially expressed during different stages of the borrelial enzootic life-cycle but they also can be shuttled to and from the surface of this organism at different points during infection (Hefty et al., 2002b). The fact that many of the lipoproteins studied to date are located in the periplasm or not surface-exposed during mammalian infection precludes specific antibodies from helping to affect clearance of the organism. Therefore, it has become of utmost importance to fully define the expression patterns of candidate surface proteins and fully delineate their cellular location during mammalian infection. At this time, it is not entirely clear how lipoproteins are retained in the periplasm and/or shuttled to the cell surface. While the B. burgdorferi genome encodes the necessary machinery for Sec translocation across the inner membrane (Fraser et al., 1997), it has been proposed that Borrelia may utilize a distinct pathway for lipoprotein transport from the periplasm to the surface of the outer membrane (Schulze and Zuckert, 2006).

The genetic makeup of B. burgdorferi is quite unusual in that it contains a linear chromosome and numerous linear and circular plasmid elements (21 plasmids in total) (Fraser et al., 1997;Casjens et al., 2000). Although the B. burgdorferi chromosome is rather small (approximately one megabase), the complexity and large sizes of many of the plasmids (some larger than 50 kb) greatly expand the DNA coding capacity of this spirochete. At the same time, however, it is currently poorly understood what role surface proteins encoded by genes on the various plasmids contribute to virulence and/or disease pathogenesis. The data accumulated thus far overwhelmingly support the hypothesis that plasmid-encoded proteins are important in Lyme disease pathogenesis and could encode antigens that are important virulence factors and/or potential vaccinogens for Lyme disease. Given that the first vaccine developed for Lyme disease was generated against the fairly well conserved, plasmid-encoded OspA, it seems likely that the identification of another outer surface protein that is well conserved throughout borrelial genospecies would be a viable candidate for a developing a new vaccine molecule. This review outlines the outer surface proteins that have been identified thus far in various borrelial species, although the main focus is on the type strain B. burgdorferi strain B31. The outer surface proteins described below fall into two main categories, lipid-modified outer surface proteins that are anchored to the outer leaflet of the outer membrane through their lipid moieties (e.g., OspA, OspB, OspC, OspD, OspE, OspF, DbpA, DbpB, CspA, VlsE, BptA, and several others with no known function) and outer surface proteins that have one or more transmembrane domains that anchor them into the outer membrane (e.g., P13, P66, BesC, BamA, Lmp1 and BB0405). The sections following provide a detailed examination of what is currently known about outer surface lipoproteins and membrane spanning outer membrane proteins of B. burgdorferi.

Borrelia burgdorferi outer surface lipoproteins

The B. burgdorferi genome encodes several lipoproteins that are localized to the surface of B. burgdorferi (Fraser et al., 1997;Casjens et al., 2000). The surface lipoproteins of B. burgdorferi are now well-recognized as important virulence determinants. As mentioned above, because of the extracellular nature of this pathogen, surface lipoproteins play an important role in virulence, host-pathogen interactions, and in maintaining the enzootic cycle of B. burgdorferi. Several borrelial surface lipoproteins have been identified that bind host proteins and promote the adherence to host cells. For instance, B. burgdorferi lipoproteins bind host glycosaminoglycans (GAGs), decorin, and fibronectin. Furthermore, lipoproteins have been implicated in evasion of the host immune response through antigenic variation and evasion of complement deposition. The various outer surface lipoproteins identified to date perform varied functions in the tick and/or mammalian environments, but there are common themes in functionality among many of the lipoproteins. For example, some lipoproteins are important for persistence in ticks, while others are important for vector to host transmission. These various functional groupings, and the surface lipoproteins that fall into each group, are outlined below in the following sections.

Persistence within ticks

OspA/OspB

Numerous surface lipoproteins have been identified that are important in colonizing and persisting within the midgut of ticks. Outer surface proteins (Osp) A and OspB were first identified based on their antigenic properties and the observation that antibodies directed against OspA were reactive with spirochetes isolated from Lyme disease patients (Barbour et al., 1983;Barbour et al., 1984;Howe et al., 1985). OspA and OspB are surface exposed lipoproteins of 31 kDa and 34 kDa, respectively (Howe et al., 1985;Fraser et al., 1997). They are cotranscribed from a single promoter and are encoded on B. burgdorferi linear plasmid (lp) 54 (Howe et al., 1986;Barbour and Garon, 1987). OspA and OspB share a high degree of sequence and similarity (~50% sequence identity), as well as structural similarity (Bergstrom et al., 1989;Fraser et al., 1997;Li et al., 1997;Becker et al., 2005). The OspA and OspB C-terminal regions are characterized by a positively charged cleft with an adjacent cavity that is lined with hydrophobic residues (Li et al., 1997;Becker et al., 2005), and it is thought that this cavity potentially binds an unknown ligand.

The role of OspA and OspB in the infectious life cycle of B. burgdorferi has only recently been elucidated. Both OspA and OspB are expressed in the midgut of unfed ticks and are downregulated upon tick feeding (Schwan et al., 1995;Schwan and Piesman, 2000;Pal et al., 2000;Ohnishi et al., 2001;Hefty et al., 2001;Hefty et al., 2002b). The abundant expression of these two lipoproteins in the tick led to the hypothesis that OspA and OspB are essential for maintenance of the spirochete within the tick environment. Correspondingly, recombinant OspA and OspB bind tick gut extracts in vitro (Pal et al., 2000;Fikrig et al., 2004). The role of OspA and OspB in the tick was further supported by in vivo examination of these proteins. In a mutant strain lacking OspA and OspB expression, mutant organisms were transmitted from infected mice to ticks and could be detected in the bloodmeal during feeding; however, the OspA/OspB mutant was unable to colonize and survive within the tick midgut (Yang et al., 2004). Interestingly, OspA alone was sufficient to restore midgut colonization to approximately 60% of wildtype (Yang et al., 2004). It is now thought that OspA mediates the attachment of B. burgdorferi to the tick midgut by binding the midgut receptor TROSPA (Tick Receptor for OspA) (Pal et al., 2004a). OspA is evidently downregulated for spirochetes to migrate out of the tick midgut and into the salivary glands. The role of OspB was further analyzed using a mutant strain that expresses OspA but lacks OspB. The OspB deficient strain was significantly impaired in its ability to colonize and survive in the tick midgut (Neelakanta et al., 2007). Taken together, these results indicate that both OspA and OspB play a role in persistence of B. burgdorferi in the arthropod vector.

OspD

OspD was initially described by Norris and colleagues as a 28 kDa surface lipoprotein encoded on B. burgdorferi plasmid lp38 (Norris et al., 1992). OspD is downregulated in response to temperature and host signals, and OspD expression reaches its peak on the B. burgdorferi surface shortly after tick feeding and detachment (Ojaimi et al., 2003;Brooks et al., 2003;Tokarz et al., 2004;Li et al., 2007;Stewart et al., 2008). Recombinant OspD can bind tick gut extracts, suggesting that OspD is involved in adherence to the tick midgut (Li et al., 2007). The role of OspD has been examined in vivo, and OspD was not required for infection of mice by needle inoculation or tick infestation (Li et al., 2007;Stewart et al., 2008). Interestingly, at least one report indicates a defect in colonization of the tick midgut by the OspD mutant strain, but this defect did not interfere with ability of the OspD mutant strain to infect naïve mice via tick infestation (Li et al., 2007). Additionally, clinical isolates have been collected that lack OspD providing further evidence that OspD is not required in the natural life cycle of B. burgdorferi (Marconi et al., 1994).

BptA

BptA (Borrelial persistence in ticks A) is encoded on plasmid lp25 by open reading frame (ORF) BBE16, and proteinase K surface accessibility assays revealed that this lipoprotein is surface exposed (Revel et al., 2005). BptA is upregulated when grown in dialysis membrane chambers that mimic the mammalian environment (Revel et al., 2002;Revel et al., 2005). A B. burgdorferi BptA mutant strain was attenuated compared to wildtype after needle inoculation of mice (Revel et al., 2005). While engorged larvae were able to acquire the BptA mutant from infected mice, the mutant spirochetes were significantly reduced in the tick midgut after molting to the nymphal stage, and no BptA mutant spirochetes were detected in tick midguts after the ticks fed to repletion (Revel et al., 2005). These data suggest that BptA is important for B. burgdorferi persistence in ticks.

Role in tick to mammalian host transmission

OspC

OspC is a 22 kDa immunodominant B. burgdorferi lipoprotein that is encoded by circular plasmid (cp) 26 (Fuchs et al., 1992;Sadziene et al., 1993;Marconi et al., 1993;Fraser et al., 1997). Although OspC has been the focus of intense research for over 15 years, the biological role of OspC in the B. burgdorferi enzootic cycle is still under investigation. To date, OspC is widely known for its reciprocal production to OspA and OspB, which has become a prototypical model for the differential gene expression that mediates spirochete transmission from the arthropod to the mammalian host (Radolf and Caimano, 2008). In the midgut of an unfed tick, Borrelia produce high levels of OspA and OspB protein, whereas OspC production is almost undetectable (Schwan and Piesman, 2000). Within 36–48 hours of a blood meal, spirochetes in the engorged tick downregulate their production of OspA and OspB, and OspC production is induced (Schwan, 2003;Schwan et al., 1995). Although there is conflicting data concerning the requirement of OspC for spirochete migration from the tick midgut to the salivary gland, and also for transmission into the host (Pal et al., 2004b;Pal et al., 2004a;Grimm et al., 2004;Ramamoorthi et al., 2005;Tilly et al., 2006), OspC has been shown to bind a tick salivary protein, Salp15, in vitro and in vivo, indicating a possible role for OspC in transmission and/or survival early during host colonization (Ramamoorthi et al., 2005). It is clear, however, that OspC is a B. burgdorferi virulence factor that is essential for infection in the murine host, since OspC deletion mutants are avirulent by both needle and tick infection routes (Grimm et al., 2004;Tilly et al., 2006). Furthermore, Rosa and co-workers demonstrated that most OspC mutants complemented in trans on a shuttle vector no longer contain the complementing plasmid shuttle vector six weeks after infection, and that OspC mutants are cleared from intradermal sites of infection within 48 hours post-inoculation (Tilly et al., 2006). These data indicate that OspC functions during very early stages of mouse infection and is not required for spirochete persistence. This conclusion is consistent with data from previous studies, which have shown that both ospC transcript and OspC protein levels are reduced within 2 weeks post infection (Liang et al., 2002a;Carroll et al., 1999;Ohnishi et al., 2001;Schwan and Piesman, 2000;Schwan et al., 1995). The mechanism of OspC function during early infection is not known, although it does not appear to involve evasion of host innate or acquired immunity, since OspC mutants are unable to infect SCID or MyD88 knockout mice (Stewart et al., 2006). Interestingly, in a recent study by Marconi and coworkers, site-directed mutagenesis of specific residues in OspC ligand-binding domain 1 (LBD1) resulted in either a loss of infectivity, or affected spirochete dissemination in mice (Earnhart et al., 2010). From these data, the authors posited that the essential function of OspC in mammalian infection is to bind an unknown host-derived ligand, which may facilitate spirochete adaptation and early dissemination within the host (Earnhart et al., 2010).

In addition to OspC function, the mechanisms by which OspC is regulated have been intensively studied. ospC expression is regulated by the Rrp2-RpoN-RpoS sigma factor cascade pathway, and is specifically dependent upon the RpoS (sigmaS or sigma38) transcription factor (Elias et al., 2000;Hübner et al., 2001;Caimano et al., 2004;Yang et al., 2005). In response to host signals during tick feeding and mammalian infection, RpoN-dependent transcription of rpoS leads to the accumulation of rpoS transcript, and in conjunction with the small RNA DsrA_Bb, RpoS expression is increased (Lybecker and Samuels, 2007;Burtnick et al., 2007;Smith et al., 2007). The RpoS subunit recognizes an extended −10 region of the OspC promoter, and direct subunit binding initiates ospC transcription (Eggers et al., 2004). ospC is just one of more than 100 genes whose expression is influenced by RpoS (Caimano et al., 2007;Ouyang et al., 2008). Interestingly, ospC gene expression is also regulated by the level of DNA supercoiling, possibly because this allows more efficient binding of RpoS to its promoter site (Alverson et al., 2003;Yang et al., 2005).

Because OspC is immunogenic during early infection and can elicit protective antibody responses (Fuchs et al., 1992;Bockenstedt et al., 1997;Gilmore, Jr. et al., 1996), OspC has been investigated as a candidate Lyme disease vaccinogen, both as a recombinant protein-based vaccine and a DNA vaccine (Earnhart and Marconi, 2007;Brown et al., 2005;Wallich et al., 2001;Scheiblhofer et al., 2003). Efforts have been complicated, however, by the fact that OspC exhibits wide sequence variation between Borrelia genospecies (Wang et al., 1999;Wilske et al., 1996;Jauris-Heipke et al., 1993), and the antibody response during infection tends to be OspC type-specific (Earnhart et al., 2005;Earnhart et al., 2007;Ivanova et al., 2009). Consequently, the numerous and different OspC genotypes will need to be included in a multi-component subunit vaccine if a broadly-protective OspC-based vaccine is to be generated.

BBA64

BBA64, also referred to as P35, is a 35 kDa B. burgdorferi antigen that is located on lp54 (Gilmore, Jr. et al., 1997;Fraser et al., 1997;Gilmore, Jr. et al., 2007). The putative BBA64 lipoprotein is membrane-anchored and surface-exposed (Brooks et al., 2006). Combined cDNA microarray and proteomic data has confirmed that BBA64 expression is increased in culture conditions that mimic the mammalian environment, such as increased temperature (37°C relative to 23°C) (Revel et al., 2002;Ojaimi et al., 2003;Tokarz et al., 2004;Brooks et al., 2006) and decreased pH (7.0 relative to 8.0) (Carroll et al., 2000;Revel et al., 2002), and also in dialysis membrane chambers (DMC) implanted into rats (Brooks et al., 2003). Additionally, BBA64 antibodies have been detected in serum from B. burgdorferi-infected mice and nonhuman primates, as well as in human Lyme sera (Gilmore, Jr. et al., 2007;Brooks et al., 2006;Gilmore, Jr. et al., 2010). Although the function of BBA64 is currently under investigation, it is becoming clear that BBA64 plays a specific role in mammalian infection. Transcript analyses determined that expression of BBA64 is detectable during tick feeding, but not detectable in replete ticks (Gilmore et al., 2001;Tokarz et al., 2004), which led to the hypothesis that BBA64 is important during tick-host transmission or during the acute stage of mammalian infection. Interestingly, Maruskova et al. demonstrated that there was no disease phenotype or alteration in virulence when mice were infected with a B. burgdorferi BBA64 null mutant (Maruskova and Seshu, 2008). However, in a more recent study by Carroll and co-workers, a B. burgdorferi BBA64 mutant was observed to be severely attenuated in its ability to infect mice when animals were challenged by the natural mode of tick infestation (Gilmore, Jr. et al., 2010). The BBA64 mutant was readily acquired by larval ticks and persisted in ticks through molting (Gilmore, Jr. et al., 2010), suggesting that BBA64 is not necessary for persistence in the tick, but is required for transmission from the tick vector to the mammalian host.

Host cell adhesion

DbpA and DbpB

Two borrelial proteins, decorin-binding proteins A and B (DbpA and DbpB), have been shown to bind host decorin (Guo et al., 1995). Decorin is a proteoglycan that consists of a protein core substituted with the GAG chains dermatan sulfate or chondroitin sulfate. Decorin interacts with collagen fibers and can be found in numerous tissues as a component of the connective tissue. Therefore, by binding decorin, DbpA and DbpB could mediate the interaction between B. burgdorferi and connective tissues. DbpA and DbpB are surface lipoproteins encoded by the dbpB/A operon (BBA24 and BBA25) located on lp54 (Hanson et al., 1998;Guo et al., 1998;Hagman et al., 1998). Both proteins are upregulated on the surface of B. burgdorferi organisms grown at reduced pH and by a temperature shift from 23° to 37°C, which suggests an important role for these proteins in the mammalian environment (Carroll et al., 2000;Ojaimi et al., 2003;Revel et al., 2002). The importance of DbpA/B in GAG binding was demonstrated by expressing DbpA or DbpB in the B. burgdorferi strain B314, an avirulent strain lacking lp54. The B314 strain was able to bind mammalian epithelial 293 cells only when DbpA or DbpB were expressed in this strain (Fischer et al., 2003).

Many studies have investigated the role of DbpA/B and decorin binding in the life-cycle of B. burgdorferi. Brown and colleagues have demonstrated the importance of B. burgdorferi decorin binding in decorin deficient mice (Brown et al., 2001). Bacterial burden in tissues of decorin deficient mice were significantly reduced as compared to wildtype mice (Brown et al., 2001;Liang et al., 2004). Needle inoculation of mice with a DbpA/DbpB deficient B. burgdorferi strain demonstrated that DbpA and DbpB are not essential for establishing an infection, but DbpA/DbpB mutant strains were attenuated in virulence (Shi et al., 2006;Shi et al., 2008;Weening et al., 2008). Despite the results from needle inoculation experiments, tick infestation studies revealed that DbpA/DbpB deficient spirochetes were able to infect mice (Blevins et al., 2008). Collectively, these experiments suggest that DbpA and DbpB play a critical role in later stages of disease, such as during dissemination and establishing a long-term chronic infection in decorin-rich tissues, but that DbpA and DbpB are likely not essential for establishing an infection in mammals.

BBK32

B. burgdorferi can bind fibronectin, a host glycoprotein that exists as a soluble serum protein or as a component of the extracellular matrix (ECM) (Szczepanski et al., 1990). Studies have shown that the B. burgdorferi protein BBK32, a 47 kDa protein encoded on lp36, can bind fibronectin and is thought to play an important role in the B. burgdorferi-fibronectin interaction (Probert and Johnson, 1998). The interaction between B. burgdorferi and fibronectin can be disrupted by pre-incubating fibronectin with BBK32 (Probert and Johnson, 1998). Furthermore, when expressed in a non-adhering B. burgdorferi strain, BBK32 was sufficient to confer binding to fibronectin and mammalian cells (Fischer et al., 2006). Further supporting the role of BBK32 as an adhesin, BBK32 is surface exposed and upregulated during tick feeding and mammalian infection (Probert and Johnson, 1998;Fikrig et al., 2000;Li et al., 2006;He et al., 2007). The interaction of BBK32 and fibronectin can be mapped to the collagen binding domain of fibronectin and a 32 amino acid stretch in BBK32 that is required for fibronectin binding (Probert and Johnson, 1998;Probert et al., 2001). In addition to binding fibronectin, it has also been shown that BBK32 can bind the host GAGs heparin and dermatan sulfate (Fischer et al., 2006). BBK32 has also been implicated in initiating the interaction of B. burgdorferi with the microvasculature in an infected mouse, which was visualized in real-time using intravital microscopy (Norman et al., 2008).

Inactivation of BBK32 in a virulent strain of B. burgdorferi revealed that the BBK32 mutant did not bind fibronectin or mouse fibroblasts cells as well as the wildtype strain (Seshu et al., 2006). The BBK32 mutant was also attenuated in its ability to infect mice via needle inoculation (Seshu et al., 2006). Nevertheless, Li et al. demonstrated that BBK32 was not essential for infection of mice in the tick-mouse model of Lyme disease (Li et al., 2006). Given that B. burgdorferi likely expresses multiple host cell adhesins, however, it is possible that BBK32 enhances dissemination in the infected host, even though no obvious phenotype was observed in the BBK32 mutant strain.

OspF

ospF was first identified downstream of the ospE gene (see CRASP section below) in a plasmid-encoded operon of B. burgdorferi strain N40 (Lam et al., 1994). Interestingly, while ospF in strain N40 is linked to the ospE gene and they are co-transcribed genes, this is unique to strain N40. The ospE and ospF genes in all other strains studied to date encode OspE and OspF on different plasmids. While OspF has not been fully characterized at the functional level, it was identified as a potential adhesin to heart tissue using an in vivo phage display system (Antonara et al., 2007). While this observation has not been further characterized, it is interesting that this protein is upregulated during mammalian infection and could be important in tissue tropism during mammalian infection (Antonara et al., 2007;Stevenson et al., 1998;Miller et al., 2000;Gilmore et al., 2001;Miller et al., 2003;Hefty et al., 2001;Hefty et al., 2002b;Hefty et al., 2002a).

Evasion of the host immune response

VlsE

B. burgdorferi is able to persist in patients for extended periods and establish chronic infection in host tissues. To evade destruction by the host immune system, the spirochete has developed evasion strategies such as antigenic variation of surface proteins. Zhang and co-workers first described antigenic variation of a 35 kDa surface lipoprotein in B. burgdorferi which they termed VlsE (variable major protein-like sequence) (Zhang et al., 1997). VlsE is similar to the well characterized variable major protein (Vmp) of the relapsing fever Borrelia (Barbour, 1993). The vlsE locus is encoded on the lp28-1 plasmid and consists of the vlsE expression site and 15 silent cassettes (Zhang et al., 1997). Within each silent cassette, there are six variable regions (VR-I through VR-VI) and six highly conserved regions. Importantly, the VlsE regions of variability are located on the membrane distal portion of the protein, which is more likely to come in contact with antibody during mammalian infection (Eicken et al., 2001).

During mammalian infection, regions of the expressed vlsE cassette are replaced with regions of the silent cassettes through a gene conversion mechanism that can result in numerous vlsE sequence products (Zhang et al., 1997;Zhang and Norris, 1998a;Zhang and Norris, 1998b). Sequence variation occurs in all six of the variable regions of the expression site, but the sequence of the silent cassettes are conserved (Zhang and Norris, 1998a;Zhang et al., 1997;Zhang and Norris, 1998b). In mice, variability of vlsE is observed as early as four days post-infection (Zhang and Norris, 1998b). These changes continue during the duration of the infection and occur at greater frequencies at later time points post-infection (Zhang and Norris, 1998b). Interestingly, clonal populations of B. burgdorferi grown in vitro or maintained within ticks retain the parental vlsE sequence, and sequence variation in immunocompetent mice occurred at a greater rate as compared to variation of vlsE in SCID mice (Zhang and Norris, 1998b). These data suggests that conversion is dependent on mammalian factors and that selection of vlsE variants occurs in the presence of an intact immune response (Indest et al., 2001;Zhang et al., 1997;Zhang and Norris, 1998b).

Presence of lp28-1, the vlsE encoding plasmid, is correlated with an intermediate infectivity phenotype of B. burgdorferi in which the spirochetes are unable to persist in tissues (Labandeira-Rey and Skare, 2001;Purser and Norris, 2000). However, strains lacking lp28-1 are able to infect and persist in SCID mice suggesting that lp28-1 is required for B. burgdorferi to survive in the presence of an intact immune system (Labandeira-Rey et al., 2003;Purser et al., 2003). A B. burgdorferi strain lacking vlsE expression was developed by deleting the region encoding this locus (Bankhead and Chaconas, 2007). Importantly, the VlsE mutant strain demonstrated a phenotype similar to an lp28-1 deficient B. burgdorferi strain. The combined data suggest VlsE as an important virulence determinant of B. burgdorferi.

Complement Regulator-Acquiring Surface Proteins (CRASPs)

Several B. burgdorferi surface lipoproteins have been identified that can bind the soluble host serum proteins factor H and/or factor H-like protein-1 (FH/FHL-1) (Hellwage et al., 2001;Kraiczy et al., 2004;Hartmann et al., 2006). Given that FH/FHL-1 are negative regulators of complement, it is thought that B. burgdorferi can evade complement mediated lysis by binding FH/FHL-1 on the bacterial cell surface. Binding of FH/FHL-1 on the B. burgdorferi surface promotes evasion of the alternative pathway of complement, and thus promotes the survival of the organism in the mammalian host. Collectively, the FH/FHL-1 binding proteins expressed by B. burgdorferi are referred to as complement regulator-acquiring surface proteins (CRASPs), and these proteins include the OspE-related proteins, CspA, and CspZ (Hellwage et al., 2001;Kraiczy et al., 2004;Hartmann et al., 2006).

The first FH-binding protein identified was the surface lipoprotein OspE (Lam et al., 1994;Hellwage et al., 2001). Hellwage et al. made the initial observation that FH/FHL-1 could be detected on the B. burgdorferi cell surface (Hellwage et al., 2001) and that the known outer surface lipoprotein OspE could interact with FH, which was demonstrated by surface plasmon resonance (Hellwage et al., 2001). The OspE-related proteins have also been referred to as Erps and Crasp-3, -4, and -5 (Stevenson et al., 1996;Kraiczy et al., 2001). OspE expression is upregulated by elevated temperature in vitro and during tick feeding and mammalian infection (Stevenson et al., 1995;Hefty et al., 2001;Hefty et al., 2002b). Many B. burgdorferi strains encode multiple OspE-related proteins that bind FH (Alitalo et al., 2002). For instance, the B. burgdorferi strain B31 encodes three OspE-related proteins. These proteins are encoded on different 32 kb circular plasmids (cp32s) by ORFs bbl39, bbp38, and bbn38 (Fraser et al., 1997;Casjens et al., 2000). bbl39 and bbp38 are 100% identical in nucleotide sequence and approximately 80% identical to bbn38 (Casjens et al., 2000). The OspE lipoproteins bind the C-terminal short consensus repeats (SCR) of FH (Alitalo et al., 2004); however, the OspE domain important in FH binding has not been fully elucidated. In fact, both N-terminal and C-terminal OspE truncations abolish FH binding, suggesting that binding to FH is discontinuous and likely dependent on a higher-ordered conformation of OspE (Alitalo et al., 2002;Metts et al., 2003;McDowell et al., 2004). In addition to FH-binding, OspE also binds host plasminogen at a distinct site from the FH binding region, and it has been suggested that this interaction may promote spirochete dissemination (Brissette et al., 2009). It is still unclear what role the binding activity of OspE may play in B. burgdorferi virulence and/or Lyme disease pathogenesis.

CspA (previously referred to as CRASP-1) was first identified as a FH-binding protein when a B. burgdorferi genomic expression library was screened for clones that could bind FH/FHL-1 (Kraiczy et al., 2004). CspA is a 27 kDa surface-localized lipoprotein encoded by ORF bba68 on lp54 (Fraser et al., 1997;Casjens et al., 2000;Kraiczy et al., 2004;Brooks et al., 2005). CspA is downregulated or completely turned off in the mammalian host environment as shown by cultivation in dialysis membrane chambers and by incubation of B. burgdorferi in the presence of human blood (Brooks et al., 2003;Tokarz et al., 2004). These observations also are consistent with the results of several studies showing that CspA is not expressed during mammalian infection (or is expressed at a dramatically low level) (Brooks et al., 2003;Tokarz et al., 2004;McDowell et al., 2006;Bykowski et al., 2007). Therefore, CspA may be most relevant in serum resistance in the tick vector during the initial bloodmeal. The interaction between FH/FHL-1 and CspA has been mapped to SCR5-7 of FH/FHL-1 (Kraiczy et al., 2004). The C-terminal 11 amino acids of CspA are required for binding to FH/FHL-1 (Kraiczy 2004). However, when the CspA crystal structure was solved, it was determined that CspA forms a homodimer and that the C-terminus is important in the interaction of the two CspA molecules (Cordes et al., 2005). Therefore, it is possible that the C-terminus plays an indirect role in FH/FHL-1 binding by stabilizing the homodimer. In fact, when the coiled coil domains of CspA are disrupted, CspA no longer binds FH/FHL-1, leading to the conclusion that binding of FH/FHL-1 to CspA requires tertiary or quaternary level folding (McDowell et al., 2005). When CspA was inactivated in B. burgdorferi, CspA was shown to be essential for serum resistance in vitro, for binding FH to the borrelial surface, and for evading deposition of complement proteins on the bacterial surface (Brooks et al., 2005;Kenedy et al., 2009). While in vitro data suggest that CspA is relevant in survival of B. burgdorferi in the presence of serum, the role of CspA in the animal model of Lyme disease has not yet been elucidated.

CspZ (previously referred to as CRASP-2) is a 27 kDa lipoprotein that has also been identified as a FH-binding protein (Hartmann et al., 2006). CspZ is encoded by ORF bbh06 on plasmid lp28-3. CspZ interacts with the SCR6-7 domain of FH/FHL-1 (Fraser et al., 1997;Casjens et al., 2000;Hartmann et al., 2006). Whether CspZ is located on the surface of B. burgdorferi is unclear. While CspZ has been detected on the borrelial surface by indirect immunofluorescence, digestion of surface proteins with proteinase K does not degrade CspZ (Hartmann et al., 2006;Coleman et al., 2008). When expressed in the serum sensitive B. burgdorferi B313 strain, CspZ enhances resistance to serum (Hartmann et al., 2006). Animal studies indicate that CspZ is expressed during mammalian infection; however, CspZ is not essential for infection of mice via tick infestation (Coleman et al., 2008). To date, CspZ is the only B. burgdorferi FH-binding protein that has been investigated in vivo.

It is now widely accepted that the two major FH/FHL-1-binding surface proteins that are most relevant to mammalian dissemination, virulence, immune evasion, and disease persistence are CspA and the OspE-related proteins. While four other surface lipoproteins encoded on various cp32 plasmids (i.e., ErpG, ErpL, ErpX, and ErpY) have been shown to bind FH/FHL-1 from other animal sources, such as cattle, cat, or dog (Stevenson et al., 2002), it is not clear what, if any, role this may play in the enzootic cycle of B. burgdorferi.

Surface lipoproteins with unknown function

In addition to the lipoproteins discussed in the preceding sections, there have also been several lipoproteins identified on the surface of B. burgdorferi that currently have no known function. Many of these were identified by Carroll and co-workers (i.e., lipoproteins BBA65, BBA66, BBA71, and BBA73) (Hughes et al., 2008) and through an examination of genes regulated by environmental cues through global expression profile analyses by Brooks et al. (Brooks et al., 2006) (BBA689, BBA36, BBA66, BBA69, and BBI42). Given their cellular location on the surface, these lipoproteins likely perform an important role in either the tick or mammalian host environment, but future studies are needed to fully elucidate their functional role(s) in B. burgdorferi virulence and/or Lyme disease pathogenesis.

Integral Outer Membrane Proteins

In addition to the numerous outer surface lipoproteins described above, B. burgdorferi also contains integral outer membrane proteins (OMPs) that have transmembrane spanning domains. OMPs are structurally different than lipoproteins, in that they do not contain N-terminal lipid anchors. Bacterial OMPs, in general, provide an array of important functions, such as nutrient acquisition (e.g., porins), antibiotic resistance (e.g., drug efflux pumps), protein transport and assembly, and cellular adhesion (Koebnik et al., 2000;Schulz, 2002;Bos et al., 2007). Likewise, B. burgdorferi OMPs also provide critical physiological functions for the spirochete cell, which is in accordance with the observation that nearly all known B. burgdorferi OMPs are encoded from stable chromosomal loci (Fraser et al., 1997). Interestingly, freeze-fracture electron microscopy has demonstrated that B. burgdorferi possesses a characteristically low abundance of integral OMPs, approximately 10-fold fewer than that detected in the E. coli OM (Lugtenberg and van Alphen, 1983;Radolf et al., 1994). This paucity of integral membrane-spanning surface proteins, combined with the apparent limited antigenicity of OMPs, has seriously hindered identification of B. burgdorferi OMPs. As a result, relatively few non-lipoprotein surface proteins have been identified in B. burgdorferi, and even fewer have been fully characterized at the functional level.

Putative porin-like proteins

P66

P66, encoded by ORF bb0603, was first identified as a 66 kDa chromosomally encoded B. burgdorferi antigen (Barbour et al., 1984;Coleman and Benach, 1987) with an immunogenic surface-exposed loop region (Probert et al., 1995;Bunikis et al., 1995;Bunikis et al., 1996). Skare and co-workers later verified P66 to be an integral membrane porin after liposome-reconstituted P66 displayed channel-forming activity in planar lipid bilayer assays (Skare et al., 1997). From this study, it was determined that P66 is a voltage-dependent, nonspecific porin with a single channel conductance measuring at 9.6 nS in 1 M KCl, which is indicative of very large 2.6 nm pores (Skare et al., 1997). P66 orthologs from other Borrelia spp. display similar biophysical characteristics, suggesting that both Lyme disease and relapsing fever spirochetes possess functional P66 orthologs (Barcena-Uribarri et al., 2010). P66 has also been shown to function as an adhesin that binds the mammalian cell receptors, β3 chain and β1 chain integrins (Coburn et al., 1999;Defoe and Coburn, 2001;Coburn and Cugini, 2003). It was further demonstrated that β3 integrin binding was mediated by a central region of the P66 protein (residues 142–384) (Coburn et al., 1999), and that a single peptide heptamer within this 242-residue region was sufficient for inhibiting attachment of B. burgdorferi to α_IIbβ3 integrins (Defoe and Coburn, 2001). Additional verification of P66 as a β3 integrin ligand was also provided by in vivo phage display experiments (Antonara et al., 2007). The virulence-associated cell adhesion properties of P66, in addition to its immunogenicity, have created an intense interest in P66 as a potential Lyme disease vaccine candidate. Interestingly, indirect immunofluorescence assays (IFA) and cDNA microarray data has demonstrated that P66 is upregulated in fed ticks and in the mammalian host, but not in unfed ticks (Brooks et al., 2003;Cugini et al., 2003), suggesting that B. burgdorferi specifically upregulates expression of the protein to aid in host cell attachment and/or tissue dissemination during mammalian infection.

P13

The chromosomal P13 protein, which is encoded by ORF bb0034, is a 13 kDa surface antigen first identified in B. burgdorferi strain B313. Strain B313 lacks almost all linear plasmids, which encode a majority of the B. burgdorferi outer surface lipoproteins (Sadziene et al., 1995). Anti-P13 monoclonal antibodies inhibited growth of strain B313 but not wild-type B. burgdorferi cells, suggesting that the abundant outer surface lipoproteins expressed by the linear plasmids in wildtype B. burgdorferi masked P13 epitopes and probably interfered with earlier identification of this integral outer membrane protein (Sadziene et al., 1995). Sequence analysis and epitope mapping indicated that P13 is a membrane-integrated protein with three transmembrane regions and a surface-exposed immunogenic loop (Noppa et al., 2001;Pinne et al., 2004). Additionally, combined results from mass spectrometry (MS), in vitro translation, as well as N- and C-terminal amino acid sequencing strongly indicated that P13 is post-translationally processed at both termini, with an N-terminal modification and a C-terminal 28-residue cleavage (Noppa et al., 2001). MS analysis confirmed that the P13 N-terminus was modified by pyroglutamination (Nilsson et al., 2002), and studies using a B. burgdorferi CptA (carboxyl-terminal protease A) deletion mutant indicated that the C-terminal cleavage was likely a result of CptA proteolysis (Ostberg et al., 2004). P13 porin activity was detected using planar lipid bilayer assays, from which it was determined that P13 possesses cation-selective pores with a single channel conductance of 3.5 nS in 1 M KCl (Ostberg et al., 2002). This channel-forming activity was eliminated in a P13-deficient B. burgdorferi mutant (Ostberg et al., 2002). Unlike P66, however, P13 is not known to be associated with virulence-related functions, and its expression has not been reported to be regulated by temperature or mammalian host-specific signals. Interestingly, P13 is a member of a B. burgdorferi paralogous gene family, which contains eight additional plasmid-encoded P13 paralogs (Fraser et al., 1997;Noppa et al., 2001;Pinne et al., 2004). One of these paralogs, BBA01, displays channel-forming properties similar to the chromosomally-encoded P13 protein (Pinne et al., 2004;Pinne et al., 2006). Furthermore, loss of the 3.5 nS membrane conductance from a p13-null mutant was restored by complementation with BBA01, suggesting that these proteins are possibly redundant at the functional level (Pinne et al., 2006).

Although P13 and P66 have been verified to possess channel-forming activity characteristic of known bacterial porins, neither protein is structurally well characterized, and both P13 and P66 have been suggested to form atypical porin structures (Noppa et al., 2001;Bunikis et al., 1995;Pinne et al., 2004). P13 is predicted to span the OM by transmembrane α-helices, which is contrary to the amphipathic β-sheet-containing beta-barrel secondary structure typical of enteric Gram-negative proteobacterial porins (Schulz, 2002;Koebnik et al., 2000). Initially, P66 was also thought to span the membrane by two α-helical transmembrane domains (Bunikis et al., 1995), although recent sequence analyses suggest that P66 may in fact form a 20–22-stranded β-barrel structure (Barcena-Uribarri et al., 2010). Future crystallography studies will be needed to fully delineate the P13 and P66 protein structures.

Another B. burgdorferi protein termed Oms28, which is encoded by ORF bba74, was originally reported to be OM-localized and to exhibit channel-forming activity (Skare et al., 1995;Skare et al., 1996). Additionally, Cluss and coworkers demonstrated that this protein was secreted from borrelial cells (Cluss et al., 2004). However, more recent biophysical and cellular localization data have suggested that BBA74 is a membrane-associated periplasmic protein that contains no integral membrane domains (Mulay et al., 2007).

Other Integral OMPs

Lmp1

Surface-located membrane protein 1 (Lmp1), encoded by ORF bb0210, is a chromosomally-encoded B. burgdorferi protein whose function, although still under investigation, may involve protection from host-adaptive immunity. Lmp1 was originally identified as a candidate adhesin selected from heart, joint, and bladder clone pools through in vivo phage display (Antonara et al., 2007). Subsequently, Pal and co-workers demonstrated that Lmp1-deficient spirochetes were severely defective in their ability to persist in murine tissues, especially in the heart, and that Lmp1 deficiency increased B. burgdorferi susceptibility to the bactericidal effects of immune sera in vitro (Yang et al., 2009). Interestingly, Lmp1 mutants survived and persisted in SCID murine tissues, suggesting that Lmp1 is needed to help B. burgdorferi resist or evade the host adaptive immune response (Yang et al., 2009). Lmp1 is a relatively large, 128 kDa surface-exposed protein predicted to contain three distinct domains of similar length: an N-terminal region (Lmp1-N) with no known conserved structural motifs, a middle domain (Lmp1-M) containing seven unique 54-residue repeats, and a C-terminal domain (Lmp1-C) rich in tetratricopeptide (TPR) repeats (Yang et al., 2009). Preliminary studies indicate that the membrane-imbedded region is contained in the N-terminal domain, and in comparison with Lmp1-M and Lmp1-C domains, the immunogenic Lmp1-N domain may be most important for spirochete survival in the murine host (Yang et al., 2010). The functions of the other two Lmp1 domains are currently not well understood, and the significance of the unique Lmp1-M repeats and of the Lmp1-C TPRs is unclear. TPR structures are ubiquitous in prokaryotic and eukaryotic proteins, and they are specifically involved in protein-protein interactions (Sikorski et al., 1990;D'Andrea and Regan, 2003). Interestingly, IFA data suggest that Lmp1-C, in addition to Lmp1-N, is surface-exposed, suggesting that the C-terminal TPRs may be interacting with host proteins at the B. burgdorferi surface to aid in spirochete survival during mammalian infection.

BesC

In silico analyses identified BesC (Borrelia efflux system protein C) as a chromosomally-encoded ortholog of the E. coli OM channel protein TolC (Bunikis et al., 2008). Protein products of besC (ORF bb0142) and the co-transcribed upstream genes besA (bb0141) and besB (bb0140), are predicted to form a bacterial resistance-nodulation-division (RND)-type protein export system known to be involved in multi-drug resistance (Yen et al., 2002;Nikaido, 2003). RND complexes are composed of three protein components: an inner membrane (IM)-localized antiporter protein, a periplasmic membrane fusion protein (MFP), and an OM channel protein, also known as OM factor (OMF) (Yen et al., 2002;Nikaido, 2003;Nikaido and Takatsuka, 2009). Bunikis et al.. demonstrated that B. burgdorferi BesC deletion mutants were 2- to 64-fold more sensitive than the wild type strain to various antimicrobial agents when tested for susceptibility in vitro (Bunikis et al., 2008). Additionally, BesC was found to possess channel-forming activity, with a large conductance of 11 nS in 1 M KCl (Bunikis et al., 2008). These studies demonstrate that BesC is a functional OMF ortholog, and in conjunction with a three-dimensional structural model of the three-component complex, the data provided by Bergstrom and co-workers indicate that BesA/B/C likely form a RND-type multi-drug efflux system in B. burgdorferi. The authors further speculated that the action of this BesA/B/C complex could account for some of the antimicrobial resistance and subsequent relapses in antibiotic-treated Lyme disease patients (Bunikis et al., 2008). Interestingly, it was observed that BesC deletion mutants were unable to establish infection in mice, suggesting that BesC may also be important for infection or for survival in the host (Bunikis et al., 2008).

BamA

BamA, which is encoded by ORF bb0795, is the B. burgdorferi outer membrane protein ortholog of the β-barrel assembly machine (BAM) (Lenhart and Akins, 2010), which is found in all diderm (dual-membraned) bacteria (Knowles et al., 2009;Voulhoux and Tommassen, 2004;Gentle et al., 2005). BamA orthologs are evolutionarily conserved, essential proteins that also have been characterized in dual-membraned eukaryotic organelles such as chloroplasts and mitochondria (Gentle et al., 2005;Voulhoux and Tommassen, 2004;Knowles et al., 2009;Gentle et al., 2004). BamA proteins in bacteria are central components of a multi-protein OM complex, which functions to assemble and localize β-barrel-containing integral OMPs into the bacterial OM (Knowles et al., 2009;Wu et al., 2005;Sklar et al., 2007). Structural characterization of B. burgdorferi BamA indicated that the 94 kDa protein contained five N-terminal polypeptide-transport associated (POTRA) structural repeats, followed by a C-terminal β-barrel region (Lenhart and Akins, 2010). Cellular localization data demonstrated that BamA is membrane-integrated, with periplasmic POTRA domains and a surface-exposed C-terminus (Lenhart and Akins, 2010). Functional assays with an IPTG-regulatable bamA gene confirmed that BamA is essential in B. burgdorferi and that depletion of BamA results in a severe decrease in the amount of integral OMPs that are efficiently exported to the borrelial surface (Lenhart and Akins, 2010). Surprisingly, BamA depletion also results in decreased levels of surface lipoproteins in the B. burgdorferi OM. It has been suggested, however, that this latter phenotype is an indirect effect of BamA depletion, perhaps due to the loss of BamA-dependent insertion of a specific integral OMP that is required for localizing lipoproteins to the surface of B. burgdorferi (Lenhart and Akins, 2010). Additionally, the B. burgdorferi BamA protein exists as an OM multi-protein complex that contains at least two other periplamsic accessory lipoproteins, BB0324 and BB0028, that interact with BamA (T. Lenhart and D. Akins, unpublished data).

BB0405

BB0405 is a 22 kDa protein whose expression and cellular localization has been relatively well described, but whose function in B. burgdorferi is currently not known. bb0405 was identified from B. burgdorferi whole genome microarray analyses as a gene that was upregulated by temperature and that also encodes a putative N-terminal signal peptide (Ojaimi et al., 2003;Brooks et al., 2006). Brooks et al. demonstrated that BB0405 was both amphiphilic and surface-exposed, as determined by TX-114 phase partitioning and proteinase K accessibility, respectively. Additionally, bb0405 encodes a putative signal peptide with a signal peptidase I cleavage site, further suggesting BB0405 is a surface-localized transmembrane-spanning OMP. Consistent with the combined data indicating that BB0405 is a surface-exposed protein, specific anti-BB0405 antibodies were observed to be bactericidal in vitro (Brooks et al., 2006). The surface localization of BB0405 suggests that it could be an excellent candidate for future Lyme disease vaccine studies.

Bgp

Given that glycosaminoglycans (GAGs) are present on most eukaryotic cells and that B. burgdorferi can bind GAGs, B. burgdorferi likely exploits this activity to interact with several different cell types and tissues during the infectious process. The B. burgdorferi surface protein Bgp (Borrelia glycosaminoglycan-binding protein) is encoded by ORF bb0588 and has been shown to bind the GAGs heparin and dermatan sulfate on the surface of mammalian cells (Parveen and Leong, 2000). Bgp is not only found as an outer surface membrane protein, but it also has been shown to be secreted from the borrelial cell (Parveen and Leong, 2000;Cluss et al., 2004). Recombinant Bgp can agglutinate erythrocytes and inhibit the interaction of B. burgdorferi and mammalian cells (Parveen and Leong, 2000), which further suggests that Bgp plays an important role in cell adhesion. Interestingly, a Bgp null strain was not required for infection of SCID mice (Parveen et al., 2006); however, it was speculated that the lack of an observed phenotype in the animal studies was likely the result of B. burgdorferi expressing other GAG-binding proteins that compensated for the Bgp deficiency in these studies.

Conclusion

The last two decades have led to the identification of several important proteins that are located on the outer surface of B. burgdorferi. Some have been shown to be bona fide virulence factors that are needed for mammalian infection (e.g., OspC), while others have been utilized as human vaccine targets (e.g., OspA). As outlined in Figure 1, some surface proteins that have been identified are specifically expressed in the tick (e.g., OspA, OspB, CspA…), while others are upregulated during tick feeding and transmission to the mammalian host (e.g., OspC, OspE, OspF, P66…). Studies have also shown that surface-exposed lipoproteins, such as OspA, OspB, OspC, OspD, OspE, and OspF, are not only localized to the cell surface but can also be detected in the periplasmic space (Figure 1), which is likely true of other surface-exposed lipoproteins. The differential expression of surface proteins is important in the parasitic strategy of B. burgdorferi and allows the organism to adapt specifically to the tick or mammalian host environment as needed (Liang et al., 2002b). Given this finding, many laboratories have tried to identify surface proteins that are expressed during the mammalian phase of the enzootic cycle to help identify novel vaccine candidates or disease modulating therapeutics for Lyme disease (Brooks et al., 2006). While numerous outer surface lipoproteins have been identified in recent years, there has only been a paucity of integral transmembrane outer membrane proteins identified and characterized over the last 20 years. The small number of integral outer membrane proteins identified is most likely due to the low abundance of the integral outer membrane proteins as compared to outer surface lipoproteins that are typically expressed at very abundant levels. Additionally, integral outer membrane proteins appear to be poorly immunogenic as compared to lipoproteins and they also can be hidden on the surface under the abundant lipoproteins that coat the borrelial surface. Given that freeze-fracture electron microscopy has shown that numerous integral outer membrane proteins are embedded in the B. burgdorferi OM (Radolf et al., 1994), it is likely that many other outer membrane proteins have yet-to-be-identified. While it may take the advent of new technologies or methodologies to identify these scarce proteins in future studies, the integral outer membrane proteins are typically very highly conserved among different genospecies, which suggests that they may be the best candidates for developing a second-generation Lyme disease vaccine. Therefore, identification of new outer surface proteins should be a high priority in the Lyme disease field, since these studies should not only help to identify proteins that may better delineate how B. burgdorferi interacts with its tick vector, but also should help to elucidate how this spirochete is transmitted to the mammal from the tick, disseminates within the infected host, and, ultimately, evades the robust host humoral and cellular immune response leading to chronic infection.

Figure 1.