INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Borrelia burgdorferi is the causative agent of Lyme disease, a multisystemic and chronic tick-borne spirochetosis characterized by a transient initial rash and erythema migrans, followed by an inflammatory process at the level of the joints, nervous system, and heart (5). Phagocytes are the cells primarily involved in first-line host defenses against bacterial pathogens. A dense infiltrate of mature neutrophils has been found at the site of tick bites (11). However, knowledge about the interaction between B. burgdorferi and phagocytes is fragmentary. These microorganisms are internalized by human polymorphonuclear leukocytes (PMNL) and monocytes-macrophages even during the preimmune phase of infection (1, 2, 24, 29). In fact, immune opsonization of B. burgdorferi is not essential, although it facilitates its uptake (1, 2, 24). Spirochetes are engulfed by conventional or coiling phagocytosis (25) and in either case are finally found inside closed phagosomes, where structural alterations of the microorganisms are apparent on microscopical observation (2).

The bactericidal mechanisms activated during engulfment include the generation of oxygen metabolites and the release of granule proteins into the phagocytic vacuole. The primary granule enzyme myeloperoxidase (MPO) catalyzes the formation of potentially cidal active chlorine compounds. Moreover, 10 other neutrophil granule proteins and peptides with cidal activity of their own have been identified and found to differ in both molecular characteristics (size and density of positive charges) and target specificities (14).

Regarding the killing of B. burgdorferi, it is still unclear which neutrophil product(s) is toxic. Upon ingestion of the spirochete, a respiratory burst, which does not seem to be essential for the intracellular killing of borreliae, is induced (4, 24). This is suggested by the fact that both mouse macrophages partially inhibited in the capacity to generate NO and oxygen radicals and human neutrophils from patients with chronic granulomatous disease, totally unable to produce superoxide anion and derivatives, retain killing activity toward the spirochete (19, 24). These observations indicate that oxygen-independent mechanisms may be particularly relevant to the killing of B. burgdorferi. Whereas low-molecular-mass neutrophil granule peptides such as bactenecins and defensins have proved ineffective (21, 26), the possible role of other granule proteins has not yet been ascertained. Preliminary work performed in our laboratory indicated that acid extracts from human neutrophil granules kill borreliae (21), an observation that has been the basis of the studies reported here.

This work describes the identification of a primary granule protein as the only component of human neutrophil granules that is responsible on its own for the killing of B. burgdorferi. The cidal activity of this protein has been characterized in some detail.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Bacterial strains and culture. Borrelia strain BITS, belonging to the species Borrelia garinii (B. burgdorferi sensu lato), was isolated from Ixodes ricinus and subcultured every fifth day. This strain was used to identify borreliacidal activities throughout the protein purification procedures. The virulent strain Tirelli (B. burgdorferi sensu stricto) used in some experiments was isolated from an infected patient and used at passage 2. Both strains were cultured at 34°C in Barbour-Stoenner-Kelly medium (BSK) (Sigma, St. Louis, Mo.) in a standard air incubator.

Bactericidal assays. Borrelia cultures were harvested during logarithmic growth by centrifugation at 15,000 × g for 10 min, at 20°C. Pellets were washed twice with Eagle minimal essential medium (MEM; Sigma) supplemented with 26 mM sodium bicarbonate (pH 7.8) and resuspended in MEM to a final concentration of 10⁶ bacteria/ml as measured in a Petroff-Hausser counting chamber. Variable amounts of the samples being tested for killing activity were added to 10⁴ to 10⁵ spirochetes, in a final volume of 0.2 ml of MEM, and incubated in a shaking bath for 30 min, at 37°C. Borrelia killing was evaluated by the most-probable-number (MPN) method, using a statistical elaboration described by Meynell and Meynell (16). Briefly, serial 10-fold dilutions were prepared and plated into five microplate wells. The plates were sealed and incubated at 34°C for 7 days, after which the wells were examined for the presence of viable spirochetes by dark-field microscopy. Viability counts were performed in triplicate.

Oxygen-dependent bactericidal activity. Spirochetes (5 × 10³) were incubated for 30 min at 37°C in MEM (which contains 0.126 M chloride) with or without the addition of MPO (0.07 µM) and/or H₂O₂ (0.33 mM), as indicated. MPO was kindly provided by G. Zabucchi et al. (33). The incubation mixtures were tested for bacterial viability (MPN) as described above. As a positive control, E. coli were treated in the same way, and their viability was measured by determining the number of CFU.

Purification and subcellular fractionation of PMNL. PMNL were isolated from the peripheral blood of healthy donors as previously described (27). Briefly, buffy coats (Blood Bank, Ospedale Maggiore, Trieste, Italy) were subjected to dextran sedimentation, and the leukocyte-rich supernatants were centrifuged on Lymphoprep (Nyegaard, Oslo, Norway) to separate polymorphonuclear (pellet) from mononuclear (interface) cells. Pellets yielded >95% neutrophils after being subjected to hypotonic shock to eliminate remaining erythrocytes.

Identification of subcellular fractions. The different fractions were identified by determining the activity of the following marker proteins, as previously described (10): 5'-nucleotidase (AMPase) (plasma membranes), lysozyme (primary and secondary granules), and MPO (primary granules only). Additionally, purified primary and secondary granules were analyzed by Sodium dodecyl-sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and found to be only marginally cross-contaminated (Fig. 1).

View larger version (53K): [in this window] [in a new window]   FIG. 1.   SDS-PAGE profiles of secondary (lane a) and primary (lane b) neutrophil granules and of a borreliacidal primary granule acid extract (lane c). Molecular weights of markers are indicated in thousands.

Preparation of granule extracts. All procedures were carried out at 0 to 4°C. In the first stage, acid extracts from postnuclear supernatants (total extract) and from purified primary or secondary granules were prepared by overnight extraction with 0.2 M sodium acetate (pH 4.2) with continuous rotation.

Isolation of proteins from primary granule extracts. (i) Gel filtration. Primary granule acid extracts were loaded onto a Sephadex G-75 column (1.4 by 46 cm; total volume 70 ml; bromophenol blue exclusion volume = 96 ± 4 ml) equilibrated with 0.375 M sodium acetate (pH 3.9) at 4°C. Fractions of 1.5 ml were collected throughout at a flow rate of 40 ml/h.

(ii) Ion exchange. Selected fractions from gel filtration were diluted 15-fold with water and subjected to high-pressure liquid chromatography (HPLC) on Shodex SP columns (Waters, Milford, Mass.) equilibrated with 20 mM sodium acetate (pH 5.5). Elutions were performed with 0 to 1 M NaCl gradients in the same buffer, and A₂₁₄ was recorded as a function of the retention time (minutes).

SDS-PAGE. SDS-PAGE was performed with 11% resolving gels, which were stained with Coomassie blue R250.

Protein sequencing. HPLC fractions with Borrelia-killing activity, which showed a single band with an apparent molecular weight of 27,000 on SDS-PAGE (Fig. 3), were subjected to electrophoresis followed by Western blotting on polyvinylidene difluoride membranes as described by Matsudaira (15). The stained, single protein bands were excised from the membranes and directly subjected to N-terminal amino acid sequence analysis, which was performed with an Applied Biosystems pulse liquid-phase sequencer (model 477A) equipped with an on-line analyzer (model 120A) of phenylthiohydantoin derivatives of amino acids.

Proteinase activity assays. Human leukocyte elastase (trypsin-like) activity was measured by using N-methoxysuccinyl-Ala-Ala-Pro-Val p-nitroanilide (Sigma) as the chromogenic substrate. Human leukocyte cathepsin G (chymotrypsin-like) activity was measured by using N-succinyl-Ala-Ala-Pro-Phe p-nitroanilide (Sigma) as the substrate. Briefly, samples to be tested were incubated in microplate wells (total volume, 0.2 ml) with 2 mM substrate in 0.1 M HEPES (pH 7.5). After 30 min at room temperature, the A₄₀₅ was measured with a Titertek Multiscan MCC/340 (LabSystems, Helsinki, Finland).

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Effect of the MPO-H₂O₂-Cl bactericidal system. Since chlorination constitutes a crucially important oxygen-dependent microbicidal strategy employed by neutrophils (12), we tested the efficacy of such system on B. burgdorferi. Table 1 shows the results of a representative experiment (n = 3) which indicate that in the conditions used, the complete system kills E. coli very efficiently but fails to produce a decrease in the number of viable B. burgdorferi.

View this table: [in this window] [in a new window]   TABLE 1.   Effects of the MPO-H2O2-Cl system on the viability of B. burgdorferi and E. coli

Oxygen-independent killing of borreliae. As opposed to the oxygen-dependent chlorinating system, neutrophil acid extracts containing proteins from the different types of granules killed B. burgdorferi (Table 2). To establish the source of the killing activity, primary granules were separated from secondary ones by using Percoll gradients. The MPO/lysozyme ratios were 19.1 ± 2.1 and 0.17 ± 0.03 (means ± standard errors, n = 4) for the bands corresponding to primary and secondary granules, respectively, which indicated a good separation. The purity of these granule fractions was additionally assessed by determining their protein patterns on SDS-PAGE (Fig. 1, lanes a and b). Borrelia killing activity was found to be present only in acid extracts from primary granules (Fig. 1, lane c; Table 2). As MPO (a major primary granule protein) was not microbicidal towards B. burgdorferi (Table 1), the possibility of other primary granule components being responsible for the killing activity was examined in detail.

Separation of primary granule components and determination of their proteolytic and borreliacidal activities. Primary granule extracts were first subjected to gel filtration on Sephadex G-75. Figure 2 (representative of four separate experiments) shows the SDS-PAGE protein pattern, proteolytic activity, and borreliacidal activity of the fractions 17 to 35. None of the preceding fractions killed borreliae or had proteolytic activity. There was a good correspondence between elastase- and cathepsin G-like proteolytic activities and killing of the spirochetes.

View larger version (48K): [in this window] [in a new window]   FIG. 2.   Gel filtration (Sephadex G-75) of an acid extract from primary granules. (A) SDS-PAGE protein pattern of fractions 17 to 35. Molecular weights of markers are indicated on the left in thousands. (B) Cathepsin G-like () and elastase-like () proteolytic activities (10 µl of each fraction), expressed as the change in A405 after 20-min incubations with the specific substrates (see Materials and Methods). (C) Borreliacidal activity of fractions 17 to 35 (10 µl of each), arbitrarily expressed as follows: +, decrease in the number of bacteria of more than 1 logarithm; ±, decrease of less than 1 logarithm; , no killing.

View larger version (22K): [in this window] [in a new window]   FIG. 3.   HPLC separation of proteins from gel filtration fractions with borreliacidal activity. (A) Fractions 20 (a), 22 (b), 24 (c), and 26 (d) from the G-75 column (Fig. 2) were subjected to cation-exchange HPLC. The vertical bars indicate 0.1 optical density unit at 214 nm. The peaks with borreliacidal activity are highlighted by grey shading. (B) Electrophoretic patterns of the borreliacidal HPLC fractions indicated as shaded areas in the elution profiles shown above. Molecular weights of protein markers are indicated on the left in thousands.

Characterization of the Borrelia killing activity of elastase. HPLC-purified elastase was assayed for killing activity toward E. coli, S. faecalis, and B. burgdorferi. There was no effect on the first two bacteria, whereas B. burgdorferi was very efficiently killed (Table 4), even by doses smaller than those used for the other two bacteria (not shown). To exclude the possibility of our laboratory Borrelia strain having become more susceptible to killing, as reported by Moody et al. (20), we tested the cidal activity of our purified elastase on a freshly isolated virulent strain and obtained identical results (Table 4).

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

This work describes for the first time the identification and characterization of a B. burgdorferi-killing factor from the human neutrophil. The results demonstrate that (i) oxygen-independent mechanisms are active against these spirochetes, whereas the MPO-H₂O₂-Cl system is not effective in conditions in which E. coli is efficiently killed, and (ii) among all granule proteins, only elastase possesses Borrelia-killing activity on its own. This protein, while ineffective toward E. coli and S. faecalis, was found to be active against B. burgdorferi. The killing activity is independent of the proteolytic activity.

Human neutrophils contain in their secretory granules a number of potentially cidal proteins such as cathepsin G, azurocidin, bactericidal permeability-increasing protein, proteinase 3, and defensins (9) which either on their own or, as in the case of MPO, in conjunction with oxygen metabolites (12) are active against different microorganisms. It was somewhat unexpected that among all granule proteins, only elastase was found to kill B. burgdorferi in our experimental conditions. This protein, which consists of 218 amino acid residues and contains two asparagine-linked carbohydrate side chains (28), is a trypsin-like proteolytic enzyme without a clear-cut independent antibacterial activity (6, 31). In fact, only extremely high, arguably nonphysiological concentrations have been reported to kill Capnocytophaga sputigena (17). Nonetheless, a helper role has been attributed to elastase, which appears to potentiate the cidal effect of other active proteins. This has been observed to be the case for the killing of C. sputigena, where very high concentrations of azurocidin become cidal when combined with elastase (18). Interestingly, the interaction between these two granule proteins was found to be enzyme dependent. Similarly, synergy between elastase and MPO or cathepsin G has been reported to result in the killing of E. coli and Staphylococcus aureus, but in this case the potentiating effect was unaffected by heating and therefore unrelated to the proteolytic activity of elastase (23).

We found that elastase was borreliacidal on its own. This was independent from its proteolytic activity, in keeping with what has been reported for other members of the serprocidin family. In fact, cathepsin G kills S. aureus, S. faecalis, E. coli, Pseudomonas aeruginosa (22), and Acinetobacter sp. (30) even after being deprived of its enzymatic activity. Additionally, azurocidin is a killing protein yet is not proteolytic (32). This finding suggests that the cidal function of these proteins is unrelated to a degradative activity.

Our results show that borreliae are very sensitive to elastase. In fact, concentrations of 3 to 5 µg/ml are sufficient to kill the microorganisms in the in vitro assay. These amounts are compatible with a physiological situation in which a very small proportion (<0.001%) of the total elastase content of the neutrophil (1.5 µg per 10⁶ cells [8]) would be secreted into spirochete-containing phagocytic vacuoles (ca. 0.2 µm in diameter and 30 µm in length), assuming a minimum of one ingested microorganism per cell. Although we cannot exclude completely the possibility of a minor but very potent contaminant being responsible for the killing of borreliae, this seems unlikely in view of the fact that such a contaminant should have copurified with elastase, i.e., should be similar in molecular size and have the same charge characteristics.

Regarding the possible borreliacidal mechanism of elastase, it may be associated to its interaction with some unknown component of the outer membrane of the spirochetes. This component would be present in both laboratory and freshly isolated, virulent strains, which we have observed to be equally susceptible to elastase (Table 4). Interactions at the level of the outer surface of the microorganisms can be lethal, as demonstrated by the fact that a monoclonal antibody to OspB has been found to be borreliacidal (7). It remains to be established whether elastase acts on borreliae through such a mechanism.