|Home | About | Journals | Submit | Contact Us | Français|
As the Lyme disease spirochete Borrelia burgdorferi shuttles back and forth between arthropod vector and vertebrate host, it encounters vastly different and hostile environments. Major mechanisms contributing to the success of this pathogen throughout this complex transmission cycle are phase and antigenic variation of abundant and serotype-defining surface lipoproteins. These peripherally membrane-anchored virulence factors mediate niche-specific interactions with vector/host factors and protect the spirochete from the perils of the mammalian immune response. In this issue of Molecular Microbiology, Tilly, Bestor and Rosa redefine the roles of two lipoproteins, OspC and VlsE, during mammalian infection. Using a variety of promoter fusions in combination with a sensitive in vivo “use it or lose it” gene complementation assay, the authors demonstrate that proper sequential expression of OspC followed by VlsE indeed matters. A previously suggested general functional redundancy between these and other lipoproteins is shown to be limited and dependent on an immunodeficient experimental setting that is arguably of diminished ecological relevance. These data reinforce the notion that OspC plays a unique role during initial infection while the antigenically variant VlsE proteins allow for persistence in the mammalian host.
Spirochetal infections have had and continue to have a significant impact on human morbitity. The global spread of syphilis likely started in the 15th century when ships returning from the newly discovered Americas introduced Treponema pallidum to a thus far naïve European population and wasn’t effectively controlled until the general availability of penicillin in the 1940s (Harper et al., 2008). Some evidence suggests that leptospirosis and its agent Leptospira interrogans crossed the Atlantic in the other direction and may have shaped the history of early European settlements (Marr and Cathey, 2010), only to become the most common zoonotic disease worldwide. Similarly, clinical descriptions from the 1880s and molecular evidence from more than 5000-year-old tissue samples indicate that Borrelia burgdorferi was present in Europe long before re-emerging as the cause of an infectious syndrome on the East Coast of the United States in the 1970s (Keller et al., 2012); Lyme borreliosis today is the most common vector-borne infection in temperate climates of the Northern hemisphere.
The Machiavellian ambition and triumph of these three spirochetal agents probably best articulates itself in their common propensity to cause persistent infection, particularly in reservoir hosts that are able to efficiently spread the disease. With the exception of syphilis, humans are considered incidental, “dead-end” hosts that will not facilitate transmission of the disease but nevertheless suffer from the consequences of an untreated infection. The mechanisms that allow B. burgdorferi to persevere inside the natural tick-vertebrate transmission cycle as well as during human infection have been the focus of a large number of studies. The bacterium’s major persistence traits can be classified under three categories. First–and common to all spirochetes–is the high degree of motility that enables the slender bacteria to penetrate and disseminate in host tissues. Cases in point for motility’s importance in the transmission cycle of B. burgdorferi are that mutants in the major B. burgdorferi periplasmic flagellar protein FlaB, which lack the typical flat-wave motility, show reduced viability inside the tick and are unable to establish infection in an immunocompetent mouse model of infection (Motaleb et al., 2000; Sultan et al., 2013). The second and third categories, adhesion and immune evasion, both involve processes occurring at the host-pathogen interface. On the B. burgdorferi side, this interface is dominated by a large and diverse set of peripherally membrane-anchored surface lipoproteins. It therefore may not be surprising that an ever-increasing number of surface lipoproteins have been shown to bind host factors such as extracellular matrix components (e.g., fibronectin, proteoglycans, collagen or laminin), protease precursors (plasminogen), or complement regulatory factors (e.g., factor H) (reviewed in Radolf et al., 2012). One could expect that some promiscuity and spatiotemporal redundancy in binding these host factors be built into the system, supposedly to safeguard efficient infection in a variety of host species. Single-gene knockouts therefore may have rather subtle and time-restricted phenotypes, which can lead to erroneously absolute conclusions about a protein’s lack of role in the infectious process. Recent state of the art studies have begun to use competitive infection assays to accentuate any phenotypical differences between the wild type and a particular mutant (Bestor et al., 2012). Current best practices for infection studies include the use of infected ticks in transmission to susceptible model animals, which is most authentic and practical to obtain qualitative data. Subsequent quantitative studies employ syringe injection, where the inoculating dose (usually around 103 wild-type bacteria) can be controlled more easily. Despite careful calibration, however, the latter does not necessarily mimic an infection via tick bite, where estimated tens of already host-adapted bacteria are sufficient to initiate infection (Ohnishi et al., 2001).
B. burgdorferi cells were first caught in the act of this host adaptation when Schwan and colleagues observed phase variation of two surface lipoproteins, OspA and OspC. While OspA was abundantly expressed in the midgut of unfed ticks, it was replaced by OspC when these ticks were fed on uninfected mice. (Schwan et al., 1995). This seminal finding laid the foundation for many gene regulation studies that followed. Most recently, changes in the metabolic status have been implied to be the overarching trigger for this phase variation (Jutras et al., 2013). OspA, the sole component of a first generation “transmission-blocking” Lyme disease vaccine (Steere et al., 1998), was ultimately shown to serve as a tick-specific persistence factor by acting as a midgut adhesin (Pal et al., 2004) and by shielding the bacterium from Borrelia-specific antibodies in the incoming blood meal (Battisti et al., 2008), thereby mitigating any effects of a tick’s fortuitous feeding on a previously exposed host. Meanwhile, studies of OspC led to proposals of a variety of sometimes-contentious functions that yet converged on a role during tick-to-mammal transmission and early colonization of the mammalian skin (reviewed in Radolf et al., 2012).
Investigations into immune evasion mechanisms quickly focused on an intriguingly complex locus on one of B. burgdorferi’s linear plasmids that expressed VlsE. VlsE belongs to a family of immunodominant variable major surface lipoproteins or VMPs that were involved in multiphasic antigenic variation in related Borrelia species causing relapsing fever (Zhang et al., 1997). Similarly to the relapsing fever Borrelia VMP system, random recombination of silent vls gene cassettes into an expression site results in segmental gene conversion that permits long-term infection with B. burgdorferi in spite of a robust adaptive immune response (Coutte et al., 2009). The resistance of most Lyme disease spirochete isolates to killing by complement, a component of the innate immune system, was eventually traced back to three evolutionarily distinct surface lipoproteins that bind host-derived complement regulators: CspA, CspZ, as well as OspE and some of its related Erp proteins. Collectively called complement regulator-acquiring surface proteins (CRASPs), they bind factor H or one of its splice variants, which averts accumulation of active complement factor C3b on the bacterial surface. This preventive measure decreases opsonization and subsequent phagocytosis as well as blocks formation of the membrane attack complex that would otherwise lead to bacterial cell lysis (reviewed in Kraiczy and Stevenson, 2013).
About 7 years ago, a set of experiments suggested that OspC might play a biological role similar to the CRASPs. Rosa and colleagues showed in a well-controlled dissemination experiment that a ΔospC mutant, in contrast to the wild type or complemented strain, was cleared from an intradermal injection site in immunocompetent mice within 48 hours (Tilly et al., 2007). In a previous study, they had also shown that an ospC gene disruption mutant was unable to infect mice deficient in MyD88, an integrator of toll-like receptor (TLR) signals (Stewart et al., 2006). This excluded an involvement of the latter branch of the innate immune response and led the authors to hypothesize that OspC may inhibit phagocytosis by limiting opsonization by complement (Tilly et al., 2007). Liang and colleagues subsequently surmised that the lack of OspC could compromise the integrity of the outer membrane lipoprotein layer, thereby making the cells an easier target for the innate immune response. Supporting this hypothesis, their ospC gene disruption mutant was able to infect and disseminate in severely combined immunodeficient (SCID) mice when heterologously complemented with forcibly stabilized recombinant plasmids expressing OspA, OspE or VlsE from a constitutive promoter (Xu et al., 2008a). A follow-up study showed–maybe not surprisingly–that such dysregulated expression of OspA and VlsE significantly altered the infection’s dissemination and clearance dynamics (Xu et al., 2008b).
In this issue of Molecular Microbiology, Tilly, Bestor and Rosa (Tilly et al., 2013) revisit and expand on these findings by extending and perfecting a gene complementation and in vivo selection approach that was pioneered in an earlier study (Tilly et al., 2006). The simple but elegant system uses different promoter-lipoprotein fusions delivered via an autonomously replicating recombinant plasmid. In contrast to the system used by Liang and colleagues, the transcriptional fusions are driven by the ospC promoter to ensure appropriate spatiotemporal expression, and the recombinant plasmid is not artificially stabilized by insertion of an essential virulence determinant missing from the background strain. In this “sensitized” experimental setup, retention of a particular expression plasmid in a particular background during infection indicates–by proxy–an essential role for the expressed protein, while loss indicates that its expression is dispensable (or even detrimental) at that particular phase of the infectious process. The combination of these tools allowed the authors to tease out and further clarify the roles of OspC and VlsE during early and persistent infection. Expression of OspA or VlsE from the ospC promoter was unable to compensate for the absence of OspC in immunocompetent as well as SCID mice. Confirming previous findings by others (Lawrenz et al., 2004; Bankhead and Chaconas, 2007), the absence of VlsE impacted persistence in immunocompetent mice only. In SCID mice, however, expression of OspC was required in the absence of VlsE, as a recombinant plasmid expressing OspC from its own promoter was absolutely retained in a vlsE-ΔospC background while being readily lost in a vlsE- only background. Supporting this finding, the vlsE-ΔospC double mutant was non-infectious for SCID mice. Together, this indicates that OspC can carry out the duties of VlsE during persistent infection as long as OspC is not targeted by antibodies. As observed earlier (Xu et al., 2008b), overexpression of proteins tended to have an attenuating effect in immunocompetent mice. Two inconsistencies with previous studies remain unresolved. First, constitutive expression of VlsE was unable to rescue an OspC-deficient strain in SCID mice, which is in conflict with the data by Liang and colleagues (Xu et al., 2008a), but in agreement with the here presented ospC promoter-derived data. Second, and somewhat confounding the otherwise clear data is the authors’ discovery that the generally non-infectious phenotype of their ΔospC mutant is sporadically leaky in SCID mice. Unfortunately, the described wide-ranging efforts to identify the cause prove Sisyphean.
In conclusion, this study proves that extracting the secrets from a pathogen that has honed its ways through eons of evolution sometimes may require subtle tools that minimally disrupt subtle and sometimes covert mechanisms. The Borrelia surface has been metaphorically likened to a “rainforest” (Bunikis and Barbour, 1999), where lipoproteins may form different layers of the canopy. Introduced overrepresentation of some of them may sufficiently disturb interactions within the surface proteome to have unintended consequences like the shielding or exposure of other virulence factors at the host-pathogen interface. The basic structural and functional information and the tools are in place to initiate a comprehensive and comparative structure-function analysis of OspC and VlsE both in vitro and in vivo. The disordered N-terminal tethers peptides of OspC and two other surface lipoproteins are already known to contain surface targeting signals (Kumru et al., 2011; Schulze et al., 2010), and their variable lengths may properly position the different lipoproteins within the complex surface architecture. Mutations of a putative ligand-binding domain at the dimer interface of OspC was shown to affect both infectivity and dissemination (Earnhart et al., 2010), but it remains to be determined whether this is due to a direct loss of function in binding any ligand. The overall protein folds of the OspC dimer and the VlsE monomer are quite similar, with alpha-helical bundles presenting a dome of membrane-distal variable loops to the host (see Fig. 1). Various experimental data already support a role for these loops in protein-specific functions, particularly for VlsE (Zhang et al., 1997; Eicken et al., 2002; McDowell et al., 2002; Coutte et al., 2009; Rogovskyy and Bankhead, 2013). It is tempting to speculate that the observed partial surrogacy of OspC for VlsE is rooted in the normally well-timed reciprocal display of these two structurally related and abundant proteins at the host-pathogen interface (Fig. 2). Similar to OspA hindering the access of antibodies and proteases to an extracellular loop region of the outer membrane porin P66 (Bunikis and Barbour, 1999), OspC may “cover” for VlsE within the bacterium’s surface lipoproteome layer–but more than just in the literal sense and for as long as the host allows.
The author would like to thank Brian Stevenson for critical reading of the manuscript. Current work in the author’s laboratory is supported by NIH Grant P30 GM103326 and a University of Kansas Medical Center Research Institute Lied Basic Science Pilot Grant.