Synthesis of autoinducer 2 by the lyme disease spirochete, Borrelia burgdorferi

Abstract

Defining the metabolic capabilities and regulatory mechanisms controlling gene expression is a valuable step in understanding the pathogenic properties of infectious agents such as Borrelia burgdorferi. The present studies demonstrated that B. burgdorferi encodes functional Pfs and LuxS enzymes for the breakdown of toxic products of methylation reactions. Consistent with those observations, B. burgdorferi was shown to synthesize the end product 4,5-dihydroxy-2,3-pentanedione (DPD) during laboratory cultivation. DPD undergoes spontaneous rearrangements to produce a class of pheromones collectively named autoinducer 2 (AI-2). Addition of in vitro-synthesized DPD to cultured B. burgdorferi resulted in differential expression of a distinct subset of proteins, including the outer surface lipoprotein VlsE. Although many bacteria can utilize the other LuxS product, homocysteine, for regeneration of methionine, B. burgdorferi was found to lack such ability. It is hypothesized that B. burgdorferi produces LuxS for the express purpose of synthesizing DPD and utilizes a form of that molecule as an AI-2 pheromone to control gene expression.

FIG. 1.

Metabolic pathways found in many organisms that lead to synthesis of AI-2 and recycling of homocysteine. Me-THF, 5-methyltetrahydrofolate. DPD can spontaneously cyclize and/or combine with borate to produce at least two different, interconvertible forms of AI-2 (12, 33, 43). Many characterized organisms are capable of regenerating methionine from homocysteine by use of one or more methionine synthase enzymes, such as MetE or MetH (53). The spirochetes Treponema pallidum and T. denticola both encode homologs of Pfs but lack homologs of LuxS or methionine synthase and are thus predicted to produce SRH as a waste product and not regenerate methionine. The spirochete Leptospira interrogans can apparently complete the entire cycle, as it contains homologs of SAH hydrolase and Me-THF-dependent methionine synthase. Studies described in this report indicate that B. burgdorferi produces DPD and homocysteine through Pfs and LuxS but lacks the ability to salvage homocysteine.

FIG. 2.

Diagram of the B. burgdorferi chromosomal region containing the pfs and luxS genes. Studies described in this work demonstrated that ORF BB0375 encodes a functional Pfs enzyme. The ORF between pfs and luxS has been shown to encode a functional S-adenosylmethionine synthase (MetK) enzyme (S. P. Riley and B. Stevenson, unpublished results). These three genes appear to form an operon with ORF BB0374, a gene lacking significant homology to any previously characterized ORF.

FIG. 3.

Expression of (A) luxS and pfs genes and (B) Pfs protein by culture-grown B. burgdorferi. mRNAs were detected by RT-PCR either with (+) or without (−) added reverse transcriptase. Proteins were detected by immunoblotting using rabbit polyclonal antiserum raised against recombinant B. burgdorferi Pfs protein. For an as-yet-unknown reason, the native Pfs protein exhibits a lower mobility in polyacrylamide gel electrophoresis than is predicted by its molecular mass. Numbers on the left in panel A are sizes in base pairs, and those in panel B are molecular masses in kilodaltons.

FIG. 4.

Production of AI-2 by B. burgdorferi. Strains 297 (wild type) and AH309 (luxS) were diluted into fresh medium and grown for 1 to 7 days at 34°C. Culture densities were determined using a Petroff-Hausser counting chamber and are illustrated as a growth curve. V. harveyi reporter strain BB170 was cultured with each B. burgdorferi culture supernatant, and induced luminescence recorded. Values illustrated are those determined following 3 h of incubation. Luminescence values obtained for the negative control AH309 supernatants were subtracted from corresponding values obtained for strain 297. Statistically significant (>90% confidence interval by independent sample t test) mean luminescence values are illustrated as grey rectangles. Error bars represent standard deviations of two to five separate experiments.

FIG. 5.

Addition of in vitro-synthesized DPD affects protein expression profiles of cultured B. burgdorferi. Different isoelectric focusing and electrophoresis conditions reveal different portions of the B. burgdorferi proteome: illustrated are representative two-dimensional gels using nonlinear isoelectric focusing between pH 3 and 10. (A) A representative, complete two-dimensional gel. (B through H) Enlarged sections of two-dimensional gels corresponding with the boxed area shown in panel A. Cultures of strain 297 were incubated in plain medium (B) or in medium containing either 1 or 2 μM in vitro-synthesized DPD and homocysteine (Hcy) (C and D), Hcy alone (E), or SAH alone (F). Cultures of strain AH309 were grown in either plain medium (G) or medium containing 2 μM in vitro-synthesized DPD and Hcy (H). Signal strengths of all detected proteins were compared within each gel. Four representative proteins visible in these gels whose relative expression levels were increased by addition of DPD/Hcy are indicated by arrows. Identities of these proteins have yet to be confirmed. As would be expected, relative mobilities of proteins during the first dimension of separation (isoelectric focusing) occasionally varied somewhat between different gels. Numbers at left are molecular masses in kilodaltons.

FIG. 6.

Effects of DPD/AI-2 on B. burgdorferi VlsE protein expression. Bacteria were cultured in the presence of indicated concentrations of in vitro-synthesized DPD/AI-2 and then analyzed by immunoblotting using VlsE-directed antiserum. As controls for equal loading, membranes were also analyzed using a monoclonal antibody directed against the constitutively expressed FlaB (flagellin) protein.