Lack of a Role for Iron in the Lyme Disease Pathogen

See allHide authors and affiliations

Science  02 Jun 2000:
Vol. 288, Issue 5471, pp. 1651-1653
DOI: 10.1126/science.288.5471.1651


A fundamental tenet of microbial pathogenesis is that bacterial pathogens must overcome host iron limitation to establish a successful infection. Surprisingly, the Lyme disease pathogenBorrelia burgdorferi has bypassed this host defense by eliminating the need for iron. B. burgdorferi grew normally and did not alter gene expression in the presence of iron chelators. Furthermore, typical bacterial iron-containing proteins were not detected in cell lysates, nor were the genes encoding such proteins identified in the genome sequence. The intracellular concentration of iron in B. burgdorferi was estimated to be less than 10 atoms per cell, well below a physiologically relevant concentration.

Successful colonization of the human host by bacterial pathogens requires that bacteria overcome strict iron (Fe) limitations imparted by the host (1, 2). In humans, the amount of free Fe (∼10 18 M) (3) is well below the levels required to support the growth of most bacteria (10 6 to 10 7 M) (4). At the onset of infection, host cells increase the production and secretion of lactoferrin to limit further available Fe and inhibit bacterial growth (5). To overcome this Fe restriction, pathogenic bacteria have developed specialized systems that aid in the acquisition and assimilation of Fe. However, this does not seem to be the case forBorrelia burgdorferi.

Analysis of membranes of B. burgdorferi indicated that they lack metalloproteins commonly associated with bacterial cytoplasmic membranes (Table 1) (6). No cytochromes, respiratory proteins, or tricarboxylic acid metalloenzymes (e.g., succinate dehydrogenase) were detected in purified inner membranes. Analysis of the complete genome sequence confirmed that B. burgdorferidoes not contain genes encoding these proteins, and in fact, contains very few genes encoding metalloproteins (7). The genes encoding a superoxide dismutase (sodA) (8), a putative ferric-uptake regulatory protein (fur), and a putative neutrophil-activating protein (napA) have been identified (7). Typically, these types of proteins require Fe as a cofactor, but their metal requirements have not been determined in B. burgdorferi. For example, theB. burgdorferi SodA has >50% identity with Mn-dependent enzymes from Thermus thermophilus, Thermus aquaticus, and Bordetella pertussis (9), and cambialistic enzymes (those active with either Fe or Mn as a cofactor) from Porphyromonas gingivalis (10) andStreptococcus mutans (11). Likewise, the Fur homolog of B. burgdorferi has 54.3% similarity to theBacillus subtilis perR gene product, which requires Mn as a cofactor and regulates dps, hemA,katA, and mrgA, all of which are involved in responses to oxidative stress and metal limitation (12,13). Therefore, these B. burgdorferi proteins may be Mn-dependent rather than Fe-dependent.

Table 1

B. burgdorferi does not contain metalloproteins commonly found in bacteria. The activity of the enzyme was assayed for in B. burgdorferi strain B31 cell extracts, and the presence or absence of the enzyme was confirmed from the genome sequence. ND, not determined.

View this table:

To investigate these observations further, we examined the effect of Fe limitation on the growth of B. burgdorferi. Because B. burgdorferi is routinely grown on complex medium [Barbour-Stoenner-Kelly (BSK-II)] supplemented with 5 to 10% rabbit serum (14), it is difficult to manipulate the Fe concentration with chelators, such as 2,2′-dipyridyl (Dp) or deferoxamine mesylate (Desferal, Ds). Culturing B. burgdorferion a modified serum-free (SF) medium supplemented with Excyte (SF-E) (15) resulted in growth rates and motility similar to those of cells grown in BSK-II. In contrast, cells cultured in SF medium were nonmotile and failed to grow (Fig. 1A). Comparisons of35S-labeled proteins isolated from cells grown in BSK-II or SF-E medium and analyzed by two-dimensional nonequilibrium pH gradient electrophoresis (2D-NEPHGE) indicated that protein profiles were not altered by growth in the modified medium. The SF-E medium, which contained 1 μM of Fe as determined by inductively coupled plasma-emission mass spectroscopy (ICP-MS), was free of siderophilins (e.g., transferrin) so that extracellular Fe concentrations could be manipulated experimentally with chelators.

Figure 1

B. burgdorferi cells are able to grow normally in Fe-limited medium. (A) Comparison of growth rates of cells grown in BSK-II (▪), BSK-II without serum (SF medium) (⧫), or SF medium supplemented with Excyte (SF-E) (▵). (B) Comparison of growth rates of cells grown in BSK-II without serum (SF medium) (⧫), SF-E medium (▵), SF-E medium + 2-2′ dipyridyl (Dp) (□), and SF-E medium + Desferal (Ds) (×). (C) Comparison of growth rates of cells grown in SF-E medium ▵), SF-E medium treated with Chelex (SF-E-Clx) (⧫), SF-E-Clx medium supplemented with Fe (×) or with Mg, Mn, and Zn (○). B. burgdorferi cells were grown in SF-E medium to a density of 5 × 107 cells per milliliter. A 5-ml culture was harvested and washed twice with 10 ml of buffer containing 20 mM Hepes, 100 mM NaCl, 10 mM EDTA (pH 7.6 with Hepes buffer), and cells were suspended in 5 ml of BSKII, SF, SF-E, SF-E + 100 μM Dp (or Ds), or SF-E-Clx. These reagents were used to inoculate 45 ml of equivalent media. FeCl3, MgCl2, ZnCl2, and/or MnCl2 were added to yield a final concentration of 10 μM, and cells were incubated at 34°C. Cell numbers were determined at 24-hour intervals by dark-field microscopy. Concentrations of metals in the different media were measured by ICP-MS (17).

B. burgdorferi cells were grown in SF-E medium or SF-E medium that contained 100 μM Dp or Ds. Reduced extracellular Fe concentrations in SF-E medium with chelators had no effect on the growth rate of B. burgdorferi (Fig. 1B), unlikeEscherichia coli, which ceases to grow when extracellular Fe concentrations drop below 0.4 μM (16). This result is similar to the observation that the growth of Lactobacillus plantarum, a free-living soil bacterium that does not use Fe (16), is the same in Fe-chelated (10 μM ethylenediamine-N,N′-diacetic acid) and Fe-containing (0.6 μM) medium. B. burgdorferi cells cultured in SF-E medium treated with the metal-chelating resin Chelex (SF-E-Clx) were nonmotile and failed to grow (Fig. 1C). ICP-MS analysis of this medium indicated that Fe, Mn, and Zn were not detectable, and Mg concentrations had been reduced 100-fold (∼60 μM). Various metals were added to the medium in order to restore growth. Mn (1 μM), Zn (1 μM), and Mg (600 μM) were all required to restore growth and motility, whereas the addition of Fe (1 μM) plus Mg (600 μM) failed to do so (Fig. 1C). Thus Mn, Zn, and Mg, but not Fe, are required for growth of B. burgdorferi.

Pathogenic bacteria respond to Fe limitation by coordinately regulating the expression of Fe-uptake systems and key virulence factors. Given that B. burgdorferi has little or no Fe requirement, we predicted that protein expression would not be altered under Fe-limited conditions in this pathogen. To test this hypothesis, we compared proteins isolated from cells grown in SF-E (1 μM Fe) or SF-E-Clx supplemented with Mg, Mn, and Zn (<0.1 μM Fe) medium by 2D-NEPHGE. No reproducible differences in protein profiles were observed. Likewise, the addition of ferrous sulfate, ferric chloride, or ferrous ammonium citrate (10 μM) to SF-E-Clx had no effect on protein patterns. Indeed, Fe concentrations >10 μM inhibited growth as well as motility and cell elongation.

To determine the specific metal requirements of B. burgdorferi directly, we measured intracellular metal concentrations using ICP-MS (17). Intracellular levels of Fe, Mn, Ca, and Mg were compared with those in E. coli and Treponema denticola, two bacteria that require Fe, and in L. plantarum, which does not require Fe (Table 2). As expected, all bacteria tested contained high levels of Mg and Ca, while Mn levels in L. plantarum were consistent with those reported previously (18). Fe concentrations in cell lysates from E. coli and T. denticola were comparable, whereas Fe was below the detection limit in lysates ofL. plantarum and B. burgdorferi. On the basis of ICP-MS, the intracellular concentration of Fe in B. burgdorferi was estimated to be <50 pmol per milligram of protein, or <102 to 103 atoms of Fe per cell. The intracellular levels of Mn were higher in B. burgdorferithan in the lysates of E. coli and T. denticola. Thus, like L. plantarum, B. burgdorferiaccumulated Mn but not Fe to significant intracellular levels.

Table 2

B. burgdorferi has extremely low levels of intracellular Fe as measured by ICP-MS. E. coli strain MC4100 and B. burgdorferi strain B31 cells were grown in BSK-II or SF-E medium, whereas L. plantarum 14917 was cultured in APT medium. Cells were harvested and washed five times in 10 ml of Hepes buffer. E. coli and B. burgdorfericells were lysed by freeze-thaw in the presence of 2% SDS. L. plantarum cells were disrupted by three passages through a French pressure cell (1.1 × 105 kPa). One milliliter of each extract was assayed for metals as described (17). Protein concentrations were determined as described by Markwell et al.(21). The data represent an average of three samples from two independent experiments.

View this table:

Uptake of 59Fe and 54Mn provided a more sensitive assay for determining intracellular metal concentrations (Fig. 2A). E. coli accumulated 7.1 × 105 atoms per cell of 59Fe and 3.8 × 104 atoms per cell of 54Mn, whereasL. plantarum contained 3.5 × 106 atoms per cell of 54Mn and 59Fe did not accumulate above background levels. Like L. plantarum, B. burgdorferi contained 8.3 × 104 atoms per cell of 54Mn, but 59Fe was undetectable. The specific activity of the 59Fe used in these experiments should allow the detection of 102 to 103 atoms of 59Fe per cell. When the specific activity of the59Fe was increased (0.5 μCi/ml) so that 75% of the total Fe was in the radioactive form, 59Fe uptake in B. burgdorferi remained undetectable. This indicated that intracellular Fe levels were <10 atoms per cell, well below a physiologically relevant quantity. The 54Mn uptake experiments suggested a requirement for Mn by B. burgdorferi. Currently, little is known about transport of essential nutrients in B. burgdorferi or other spirochetes. To determine if the transport of 54Mn was energy-dependent, we used the uncoupler, carbonyl cyanidem-chlorophenylhydrazone (CCCP), in uptake assays. CCCP inhibited 54Mn accumulation by >75% (Fig. 2B), but had no effect on the background levels of 59Fe. The uptake of 54Mn by B. burgdorferi cells was linear for 1 hour.

Figure 2

B. burgdorferi accumulated54Mn, but not 59Fe, in an energy-dependent manner. (A) E. coli strain MC4100 and B. burgdorferi strain B31cells were cultured in SF-E media to a density of 1 × 108 cells per milliliter. E. coli cells were incubated at 37°C with aeration for 12 hours, whereas B. burgdorferi cells were incubated at 34°C for 48 to 72 hours. L. plantarum cells were cultured inLactobacilli MRS broth at 37°C to a density of 1 × 107 cells per milliliter. E. coli and L. plantarum cell numbers were determined by plate count, whereasB. burgdorferi cell numbers were determined as described above. Cells in each culture were harvested by centrifugation, washed twice in 10 ml of Hepes buffer, and suspended in SF-E media containing either 59Fe or 54Mn (0.16 μCi/ml each) at a concentration of 1 × 107 cells per milliliter. For determination of 54Mn and 59Fe uptake, the labeled cells were harvested, washed five times in 50 ml of Hepes buffer, and suspended in 500 μl of Hepes buffer. Incorporated radioactivity was determined from 10-, 50-, and 100-μl aliquots in 10 ml of scintillation fluid with a LS 6000 Beckman scintillation counter. (B) B. burgdorferi strain B31 cells were cultured and suspended in SF-E medium as described above. CCCP (100 μM) was added and cells were incubated at 34°C for 1 hour before adding either 59Fe or 54Mn. The cultures were incubated at 34°C for 12 hours. Uptake was determined as described above.

We have shown that B. burgdorferi is unique among pathogenic bacteria in having evolved a novel strategy to survive Fe limitation in the human host. This has been accomplished by eliminating most of the genes that encode proteins that require Fe as a cofactor and, similar to L. plantarum, by substituting Mn for Fe in the few metalloproteins found in B. burgdorferi. B. burgdorferi is an obligate parasite with a minimal genome that lacks genes encoding the enzymes for most biosynthetic pathways. A consequence of this evolution to obligate parasitism has been the elimination of the requirement for Fe. This adaptation by B. burgdorferi has been successful, and other pathogenic bacteria with limited genomes, such as T. pallidum (950 Mb) (19) and Mycoplasma pneumoniae (650 kb) (20), may have adopted similar approaches to avoid host Fe limitation.

  • * To whom correspondence should be addressed. E-mail: frankg{at}


View Abstract

Stay Connected to Science

Editor's Blog

Navigate This Article