INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Fusobacterium necrophorum is a gram-negative, pleomorphic, obligately anaerobic, rod-shaped bacterium. It frequently is associated with necrotic disease conditions of animals and is the etiological agent of Lemierre's syndrome in humans (20). Lemierre's syndrome is characterized by an oropharyngeal infection resulting in persistent high fever and disseminated metastatic abscesses, frequently including septic thrombophlebitis of the internal jugular vein (9). F. necrophorum is the causal agent of hepatic abscesses, foot rot, necrotic laryngitis (calf diphtheria), and other necrotic lesions in cattle (15, 21). Liver abscesses in feedlot cattle and foot rot in beef and dairy cattle are of significant economic importance to the cattle industry (22, 28). Historically, F. necrophorum has been classified into four biotypes or biovars: A, B, AB, and C (15). Biotype C, now called Fusobacterium pseudonecrophorum, and biotype AB are rarely encountered in liver abscesses (4). Biotypes A and B, the most frequent types encountered in hepatic abscesses, have been assigned subspecies status: F. necrophorum subsp. necrophorum and F. necrophorum subsp. funduliforme, respectively (30). F. necrophorum subsp. necrophorum is more virulent, produces more leukotoxin and hemagglutinin, and is isolated more frequently from hepatic abscesses in cattle than F. necrophorum subsp. funduliforme (22, 28).

F. necrophorum is a normal inhabitant of the gastrointestinal tracts of animals and humans. Virulence factors and pathogenic mechanisms that contribute to the transition of this otherwise commensal organism to a pathogen are poorly understood (33). A leukotoxin, endotoxin, hemolysin, hemagglutinin, and several enzymes such as DNase and proteases have been suggested as possible virulence factors (8, 33). However, several studies implicate leukotoxin, a protein cytotoxic to ruminant polymorphonuclear cells (32, 34), as the major virulence factor (8, 33). The importance of leukotoxin as a virulence factor in F. necrophorum infections is indicated by a correlation between toxin production and ability to induce abscesses in laboratory animals (5), an inability of non-leukotoxin-producing strains to induce foot abscesses in cattle following intradermal inoculation (7), and a relationship between antileukotoxin antibody titers and protection against infection in experimental challenge studies (6, 25-27).

Bacterial leukotoxins and cytotoxins generally have molecular masses of less than 200 kDa. This includes characterized leukotoxins of Mannheimia (Pasteurella) haemolytica (104,000 Da) (10), Staphylococcus aureus (38,000 and 32,000 Da) (19), and Actinobacillus actinomycetemcomitans (114,000 Da) (14) and other pore-forming toxins of gram-negative bacteria (103,000 to 198,000 Da) (29). However, leukotoxin secreted by F. necrophorum was shown to be approximately 300 kDa in size based on Sephadex column purification and sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analyses (34, 35).

To better define the molecular nature of the F. necrophorum leukotoxin, and as a first step to determining its specific role in the virulence of this bacterium, the leukotoxin gene was isolated, its nucleotide sequence was determined, and the recombinant leukotoxin was expressed in Escherichia coli.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Preparation of polyclonal antileukotoxin antiserum. Leukotoxin from F. necrophorum subsp. necrophorum strain A25 was purified using an immunoaffinity column containing antileukotoxin monoclonal antibody F7B10 (35). Affinity-purified native leukotoxin (0.5 mg) in 100 µl of phosphate-buffered saline (PBS) was homogenized with an equal volume of Freund's complete adjuvant and injected intramuscularly into rabbits. A booster dose was given on day 21 with 0.5 mg of native toxin in 100 µl of PBS homogenized with an equal volume of Freund's incomplete adjuvant. Serum samples were collected on day 42. Naturally occurring rabbit antibodies that react to E. coli proteins were removed from the antisera as follows. Cell pellets of E. coli XL1-Blue MRF' host cells grown overnight in Luria broth were sonicated in PBS and centrifuged to remove cellular debris, and the supernatant was incubated with 100-mm-diameter nitrocellulose membranes at 37°C for 3 h. The nitrocellulose membranes were then washed twice in PBS-T (0.05% Tween 20 in PBS [pH 7.2]), blocked in 2% bovine serum albumin (BSA), and washed three times again in PBS-T. Two milliliters of rabbit antileukotoxin polyclonal antiserum was diluted 10-fold in PBS-T containing 0.2% BSA and exposed to 10 changes of E. coli lysate-treated nitrocellulose membranes for 30 min each at 37°C. The resultant polyclonal antisera had minimal reactivity against E. coli proteins. Neutralizing activity of the serum, as determined by the 3-(4,5-dimethylthiazoyl-2yl)-2,5-diphenyltetrazolium bromide (MTT) dye neutralization test and the indirect enzyme-linked immunosorbent assay (ELISA) titer, was measured as described previously (32, 34, 35).

Extraction of genomic DNA from F. necrophorum and E. coli. Chromosomal DNA was extracted from highly virulent F. necrophorum subsp. necrophorum strain A25 (17) and E. coli DH5 (F 80 [lacZYA-argF] endA1 recA1 hsdR17 deoR thi-1 supE44 gyrA96 relA1) using a modification of the method described by Hull and coworkers (13). E. coli was cultured in Luria broth with shaking under aerobic conditions at 37°C, and F. necrophorum was grown overnight in a prereduced anaerobically sterilized brain heart infusion broth in serum bottles under anaerobic conditions at 39°C (12). Cell pellets were resuspended in TES buffer (25% sucrose, 50 mM Tris-HCl [pH 7.5], and 1 mM EDTA), spheroplasted with lysozyme at room temperature for 30 min, and lysed using Sarkosyl in the presence of proteinase K at 60°C for 1 h. The product was extracted with buffer-saturated phenol and chloroform, and the DNA was precipitated in 2.5 volumes of ice-cold ethanol. The DNA pellet was resuspended in TE buffer (10 mM Tris-HCl [pH 8.0] and 1 mM EDTA) and subjected to ultracentrifugation in a cesium chloride step gradient (43.5 to 60%) containing ethidium bromide (0.4-mg/ml final concentration). The chromosomal DNA band was extracted with TE buffer and CsCl-saturated isopropanol to remove ethidium bromide and dialyzed against double-distilled water. The DNA concentration and purity were checked spectrophotometrically.

Genomic library and screening. Genomic DNA of F. necrophorum A25 was digested partially with restriction endonuclease Sau3AI, and the fragments were size-fractionated in a sucrose gradient (3). Ten- to 12-kb fragments were cloned into BamHI-digested and alkaline phosphatase-treated Lambda Zap Express vector (Stratagene Corp., La Jolla, Calif.) in accordance with the manufacturer's instructions. Recombinant lambda DNA was packaged (Gigapack gold; Stratagene) and used to infect XL1-Blue MRF' host cells (Stratagene). Plaques were lifted onto the nitrocellulose membrane and screened with antileukotoxin polyclonal antiserum using a Picoblue immunoscreening kit in accordance with the manufacturer's protocol (Stratagene). Immunoreactive clones were plaque purified three times using the polyclonal antiserum. The recombinant DNA from immunoreactive clones was rescued as phagemid (pBKCMV) clones using Exassist helper phage in the E. coli XLOLR strain in accordance with the manufacturer's protocol (Stratagene).

DNA sequencing analysis. Phagemids from immunoreactive clones, purified PCR products, and plasmid subclones were sequenced using vector-specific or internal primers with a model 373A automated DNA sequencer (Applied Biosystems, Foster City, Calif.). The DNA sequences were aligned and analyzed using Sequencher (version 3.1.1; Gene Codes Corp., Ann Arbor, Mich.) and DNA Strider (version 1.2).

Inverse PCR and sequence extension. Chromosomal DNA from F. necrophorum strain A25 was digested singly with restriction endonuclease TaqI, EcoRI, DdeI, or Sau3AI. After complete digestion of the chromosomal DNA with any one of these enzymes, the products were extracted with phenol and chloroform and precipitated with ethanol. Under dilute conditions (200 ng of digested DNA in a 100-µl total volume), DNA was self-ligated using T4 DNA ligase at 16°C overnight. Ligated DNA was extracted with phenol and chloroform, precipitated with ethanol, and reconstituted in 10 µl of nuclease-free water. Two microliters of the ligated DNA was used as the template for 100-µl PCRs with forward and reverse primers designed based on the sequence obtained from previous sequencing reactions (24). The products from inverse PCR were cloned in pCR TOPO cloning vectors (pCR2.1, pCR-BluntII, and pCR4Blunt) in accordance with the manufacturer's instructions (Invitrogen Corp., San Diego, Calif.) and sequenced directly or after subcloning using vector-specific primers. Six successive inverse PCRs were carried out to reach the 3' end of the leukotoxin gene.

Creation of gene truncations. PCR with thermostable polymerase (EXTaq; Takara Corporation, Madison, Wis.) was used to amplify five overlapping regions of the leukotoxin gene ranging in size from 1.1 to 2.8 kb (11). Chromosomal DNA from F. necrophorum strain A25 was used as the template. The forward primers were designed to contain a SacI site, and the reverse primers had an XmaI site, for in-frame insertion into the His tag expression vector pQE30 (Qiagen Inc., Valencia, Calif.). Each truncated gene product overlapped with the adjacent product by at least 100 bp. One kilobase of DNA from the 3' end of the upstream open reading frame (ORF; ups) was amplified and cloned in pQE30 vector as described above. Recombinant plasmids were transformed into E. coli host strain M15 for inducible expression of proteins encoded by cloned genes under the control of the lac promoter. The five truncated leukotoxin polypeptides and the C terminus of the upstream polypeptide were purified using nickel chelation chromatography under denaturing conditions to apparent homogeneity as indicated by silver-stained SDS-PAGE gels (data not shown).

Preparation of polyclonal antiserum against the truncated leukotoxin polypeptides. New Zealand White rabbits were injected intramuscularly with the five truncated leukotoxin polypeptides or the upstream polypeptide (0.5 mg/animal) precipitated with aluminum hydroxide. A booster dose was given on day 21 (0.5 mg/animal). Serum samples were collected on days 21 and 42, and antileukotoxin titers were determined by indirect ELISA using affinity-purified native leukotoxin (35). Leukotoxin-neutralizing activities of the day 42 serum samples were determined by the MTT dye neutralization assay using 200 U of toxin (35).

Immunoblot analysis. Affinity-purified native leukotoxin, the truncated leukotoxin polypeptides and an upstream polypeptide purified over nickel columns, whole-cell lysates from bacterial clones carrying recombinant expression plasmids, and concentrated culture supernatants were resolved by SDS-PAGE (6 or 10% acrylamide) and electroblotted to nitrocellulose membranes (Bio-Rad Minigel II electrophoresis and transfer unit). A monoclonal antibody against native leukotoxin (F7B10 [35]) or polyclonal antisera raised against native leukotoxin, various truncated leukotoxins, or upstream polypeptides were used to probe the Western-blotted proteins. Goat anti-mouse or anti-rabbit immunoglobulin conjugated to alkaline phosphatase (Sigma Chemical Company, St. Louis, Mo.) was used as the secondary antibody, and the immunoreactive proteins were detected using nitroblue tetrazolium and 5-bromo-4-chloro-3-indolylphosphate as substrates.

Cloning and expression of full-length leukotoxin ORF. A 4.3-kb DNA fragment containing the 5' end of the lktA ORF up to the internal NheI restriction endonuclease recognition site was amplified from A25 chromosomal DNA. This fragment was cloned into kanamycin resistance-encoding vector pCR Blunt II TOPO. A 5.4-kb DNA fragment extending from the NheI site to the 3' end of the lktA ORF was amplified by PCR and cloned into low-copy-number spectinomycin resistance plasmid pCL1921 (18). The two resulting plasmid clones were ligated together making use of the unique NheI site present in the lktA ORF, and the transformants were selected on media containing spectinomycin (100 µg/ml) and kanamycin (21 µg/ml). The pCR Blunt II vector-specific sequences were then removed by digesting the resultant plasmid with SacI followed by ligation under dilute conditions and selection on L agar containing 100 µg of spectinomycin/ml. Thus the entire 9,726 bp of the leukotoxin ORF was cloned in low-copy-number plasmid pCL1921 to produce pSN1999. The unique XmaI site introduced into the 3' end of the ORF and the SacI site introduced into the 5' end of the reading frame were used to clone the entire lktA coding sequence in frame into expression plasmid pQE30 to give pSN2000.

Flow-cytometric analysis of leukotoxin biological activity. Bovine peripheral polymorphonuclear leukocytes (PMNs) were isolated as described previously (32, 34). Untreated cells (negative control) or those treated with either 200 U of native leukotoxin from F. necrophorum (positive control) or whole-cell lysates from clones expressing full-length recombinant leukotoxin were tested for viability (31) by flow cytometry (Facstar; Becton Dickinson Immunocytometry Systems, San Jose, Calif.). Briefly, 1 ml of bovine peripheral PMNs (9 × 10⁶ cells/ml) was incubated with various preparations of toxin for 45 min at 37°C in a chamber containing 5% CO₂. The cells were then washed twice in 2 ml of Hanks balanced salt solution (pH 7.2) and resuspended in 300 µl of HBSS. These cells were treated for 10 min in the dark at room temperature with 10 µl of 5-mg/ml propidium iodide (PI). The red fluorescence (FL-2 [585/42]) is proportional to the number of cells which have lost membrane integrity and, therefore, do not exclude the PI. Leukocyte subpopulations were displayed in a dot plot and gated according to size based on forward scatter and granularity or 90° light scatter. A region was placed around granulocytes, cells of larger size and granularity, thus excluding monocytes, and data on 10,000 gated cells were collected. The identity of the gated cells as granulocytes was indicated by indirect immunofluorescence labeling with monoclonal antibody DH59B (VMRD Inc., Pullman, Wash.), which reacts with the granulocyte-monocyte-1 receptor. Fluorescence signals displayed as a dot plot were used to determine the percentage of positive cells by quadrant statistics.

Southern blot analysis. Genomic DNA was extracted from several strains of F. necrophorum subsp. necrophorum and F. necrophorum subsp. funduliforme isolated from ruminal contents or liver abscesses (23). Chromosomal DNA was digested to completion with HaeIII, which cleaves the leukotoxin ORF once. The digested DNA was electrophoresed in a 1% agarose gel and Southern blotted onto a nitrocellulose membrane. The full-length lktA ORF cloned in pQE30 (pSN2000) was released by digestion with SacI and XmaI, and the insert DNA was gel purified, radiolabeled with [-³⁵S]dATP, and hybridized (1).

Nucleotide sequence accession number. The nucleotide sequence of F. necrophorum subsp. necrophorum strain A25 lktA has been assigned GenBank accession no. AF312861.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Cloning and nucleotide sequence of the F. necrophorum leukotoxin determinant. A Sau3A-generated genomic library of F. necrophorum strain A25 DNA was screened using rabbit polyclonal antisera raised against immunoaffinity-purified native leukotoxin, and immunoreactive clones were identified. The clones carried inserts approximately 4.6, 5.5, and 6.3 kb in length. The immunoreactive clones containing the leukotoxin ORF (designated lktA) are depicted in Fig. 1. Inverse PCR was used to extend the cloned region to allow completion of the sequence of the lktA ORF. The 11,130-bp sequence of F. necrophorum DNA contained one complete ORF and two partial ORFs. The upstream (orfB) partial ORF comprises the first 1,018 bp. The lktA ORF initiates 16 bp downstream of the orfB ochre codon. A putative ribosome-binding site (RBS) with the sequence AAGGGGGT precedes the lktA ORF. The first two bases of the RBS are the last two bases of the orfB stop codon. The leukotoxin determinant is 9,726 bp and encodes a protein of 3,241 amino acids with an overall molecular weight of 335,956. The deduced protein sequence is unusual in that it lacks cysteine residues. The protein has substantial hydrophobic character (Fig. 2) and possesses 14 regions with sufficient hydrophobic character and length to be membrane spanning (TMpred; http://dot.imgen.bcm.tmc.edu:9331/seq-search/struc-predict.html). However, this is a secreted toxin in F. necrophorum. The potential transmembrane domains may provide a clue to the mode of action of the leukotoxin on the target neutrophils.

View larger version (15K): [in this window] [in a new window]   FIG. 1.   The leukotoxin locus of F. necrophorum. Boxes, leukotoxin ORF (lktA) and its flanking putative ORFs; lines above boxes, phagemid clones (816, 101, and 611) obtained from the immunoreactive plaques in the cloning experiments; region iPCR, sequence obtained from sequencing a series of inverse PCR clones. Plasmid pSN2000 contains the entire lktA ORF. Below the boxes are the clones expressing the truncated leukotoxin polypeptides. The numbers refer to the nucleotide positions of the boundaries of each truncation relative to the 11,130-bp sequence deposited in GenBank.

View larger version (59K): [in this window] [in a new window]   FIG. 2.   Kyte-Doolittle hydropathy plots of deduced amino acid sequences from the F. necrophorum leukotoxin gene. Lines above plot, regions of the five truncated LktA polypeptides (BSBSE, SX, GAS, SH, and FINAL).

Creation of truncated leukotoxin polypeptides and characteristics of polyclonal antisera raised against them. A 3.5-kb sequence from the 5' end of the lktA gene was amplified by PCR and cloned in frame in expression vector pQE30. Induced expression of this truncated version of the leukotoxin protein with IPTG (isopropyl--D-thiogalactopyranoside) resulted in the immediate cessation of growth and lysis of the host E. coli cells. To obtain better expression of the recombinant protein and less toxicity to E. coli host cells, smaller truncations of the leukotoxin gene were constructed. The truncated polypeptides were named BSBSE, SX, GAS, SH, and FINAL starting from the N terminus and ending at the C terminus of the leukotoxin protein (Fig. 1). Each polypeptide had an overlap of at least 21 amino acids with its adjacent polypeptide. The C-terminal truncated polypeptide of the upstream protein and the polyclonal antiserum raised against it served as a negative control in our toxicity and toxin neutralization studies. Purified truncated leukotoxin and upstream polypeptides were then analyzed by Western blotting for their reactivity against polyclonal and monoclonal antisera raised against affinity-purified native leukotoxin using Western blot analysis. Antileukotoxin polyclonal antisera reacted strongly with polypeptides BSBSE, SX, and FINAL and weakly with polypeptides GAS and SH (Fig. 3A). The antileukotoxin antibody reacted with the N-terminal polypeptide, BSBSE, but not with any other truncated leukotoxin polypeptides (Fig. 3B). As expected, the UPS (encoded by the 1-kb 3' end of orfB) polypeptide did not react with polyclonal or monoclonal antileukotoxin antibodies. Polyclonal antisera raised in rabbits against each of the truncated leukotoxin polypeptides reacted strongly with the corresponding polypeptide and also the native leukotoxin (Table 1). Antibodies raised against individual truncations reacted weakly to their adjacent polypeptides because of the presence of the overlapping amino acid sequences between them (data not shown). Antiserum raised against UPS (from the upstream ORF) failed to recognize the leukotoxin.

View larger version (44K): [in this window] [in a new window]   FIG. 3.   Western blot analysis of truncated forms of purified recombinant leukotoxin protein probed with polyclonal antileukotoxin antiserum (A) and monoclonal antileukotoxin antibody (F7B10 [35]) (B). (C) Western blot of whole-cell lysates from E. coli clones expressing the full-length recombinant leukotoxin probed with the monoclonal antileukotoxin antibody. Abbreviations: MW, molecular weight markers; Lkt, affinity-purified leukotoxin from F. necrophorum; FL-I and FL-UI, full-length clone induced or uninduced, respectively, with IPTG; Super, concentrated F. necrophorum A25 culture supernatant. Arrows, positions of the reactive BSBSE band (B) and the full-length leukotoxin (C). The amount of full-length leukotoxin in the culture supernatant in panel C was insufficient to be visualized as a distinct band in this blot.

Creation of full-length recombinant leukotoxin and its toxicity to bovine peripheral blood polymorphonuclear cells. The entire leukotoxin gene (9,726 bp) was cloned into the pQE30 expression vector. Unlike certain truncated versions of the leukotoxin protein, full-length recombinant leukotoxin upon expression was not toxic to E. coli host cells. When whole-cell lysates from clones expressing full-length leukotoxin were subjected to Western blot assays, both polyclonal (not shown) and monoclonal antileukotoxin antibodies reacted to high-molecular-weight (>220-kDa) protein species (Fig. 3C). The protein was extremely unstable, as evidenced by the presence of numerous smaller-molecular-weight species, which presumably represent breakdown products. This instability was also observed with native leukotoxin that was immunoaffinity purified from F. necrophorum culture supernatants (35). Antisera raised against all the truncated leukotoxin polypeptides, including the C-terminal FINAL polypeptide, reacted to recombinant leukotoxin, suggesting that the protein may be expressed in its full length (data not shown). As expected, the antibody raised against the upstream polypeptide failed to react to the full-length recombinant leukotoxin.

View larger version (48K): [in this window] [in a new window]   FIG. 4.   Evaluation of leukotoxic activity by flow cytometry. Membrane damage was assessed by staining the cells with PI. Shown are the values obtained after counting 10,000 PMNs (stippled bars) or the lymphocyte fraction (hatched bars). Cells were untreated (control) or treated with 200 U of affinity-purified leukotoxin from F. necrophorum (native leukotoxin) or lysates of E. coli harboring expression plasmids bearing the upstream polypeptide (SN200) or the full-length lktA ORF (SN100). U and I, lysates from uninduced cultures and cultures induced with 1 mM IPTG, respectively. Induced lysates were also tested after 1:5, 1:25, and 1:125 dilutions in PBS. The results are the averages of three experiments, and the standard deviations are indicated.

View larger version (36K): [in this window] [in a new window]   FIG. 5.   Neutralization of the toxicity of the recombinant leukotoxin assessed by flow cytometry. Membrane damage was assessed by staining the cells with PI. Shown are the values obtained after counting 10,000 PMNs (stippled bars) or the lymphocyte fraction (hatched bars). Cells were treated with a lysate obtained from IPTG-induced cultures of E. coli bearing pSN2000. The lysates were untreated (control) or treated with polyclonal antiserum raised against affinity-purified leukotoxin from F. necrophorum (anti-native leukotoxin) or with polyclonal antiserum raised against the recombinant BSBSE and GAS polypeptides. The results are the averages of duplicate experiments, and the standard deviations are indicated.

Presence of the leukotoxin determinant in F. necrophorum isolates. The leukotoxin gene was cloned and sequenced from F. necrophorum subsp. necrophorum A25, a strain originally isolated from a bovine liver abscess. Southern blot hybridization of the chromosomal DNA extracted from various F. necrophorum strains of both subspecies isolated from ruminal contents or liver abscesses was carried out (23) using the leukotoxin ORF as a probe (Fig. 6). Restriction endonuclease HaeIII was used to digest the chromosomal DNA from F. necrophorum isolates. A single recognition site for this enzyme occurs 5,933 bp from the start codon in the lktA ORF. Thus, two hybridizing fragments should be present in strains carrying this gene. All strains of F. necrophorum subsp. funduliforme isolated from liver abscesses (B17, B29, and B35) or ruminal contents (RB33 and RB37) were identical in their hybridization patterns, showing two bands at approximately 7 and 8 kb each. Also, all isolates of F. necrophorum subsp. necrophorum, except A39, isolated from liver abscesses (A21 and A25) and those isolated from ruminal contents (RA13, RA15, RA16, RA18, RA26, RA28, and RA29) had identical hybridization patterns showing two bands of approximately 10 and 11 kb each. A single band of approximately 10.5 kb, presumably a doublet, hybridized to the leukotoxin gene in chromosomal DNA of strain A39 (Fig. 6, lane 4). This suggests that some heterogeneity may be present in the leukotoxin locus sequences among strains of F. necrophorum subsp. necrophorum. However, the hybridization pattern does appear to be a good indicator for subspecies determination.

View larger version (43K): [in this window] [in a new window]   FIG. 6.   Hybridization patterns of radiolabeled lktA with Southern-blotted HaeIII-digested restriction fragments of genomic DNAs from F. necrophorum subsp. necrophorum isolates from liver abscesses (lane 1, strain A21; lane 2, A25; lane 3, A39) or ruminal contents (lane 7, RA13; lane 8, RA15; lane 9, RA16; lane 10, RA18; lane 11, RA26; lane 12, RA28; lane 13, RA29) or F. necrophorum subsp. funduliforme isolates from liver abscesses (lane 4, B17; lane 5, B29; lane 6, B35) or ruminal contents (lane 14, RB33; lane 15, RB37). Strains are described in reference 23. Lane M, DNA molecular weight markers.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

F. necrophorum subsp. necrophorum is isolated more often than F. necrophorum subsp. funduliforme from necrotic abscesses (28). The strains of F. necrophorum subsp. necrophorum produce the high-molecular-weight leukotoxin in greater quantities than strains of F. necrophorum subsp. funduliforme (33). In this study, we cloned the leukotoxin gene from highly virulent F. necrophorum subsp. necrophorum strain A25. The evidence that the lktA determinant encodes the leukotoxin is as follows. (i) The ORF encodes a 336-kDa protein, a size consistent with previous studies of the toxin (34); (ii) the protein encoded by the recombinant lktA determinant is recognized by both polyclonal and monoclonal antibodies raised against purified leukotoxin from F. necrophorum; (iii) antisera raised against polypeptides from the cloned lktA determinant recognized the native toxin in Western blots; (iv) antisera raised against two of the truncated polypeptides neutralized the toxic activity of the leukotoxin; (v) the recombinant protein expressed in E. coli is relatively more toxic to bovine neutrophils than to bovine lymphocytes (these differing degrees of toxicity to neutrophils and lymphocytes are also observed with leukotoxin that was affinity purified from F. necrophorum culture supernatants); (vi) the activity of the recombinant leukotoxin is neutralized with the antisera raised against the same two polypeptides which gave neutralizing activity against the native leukotoxin.

The leukotoxin ORF is 9,726 bp long and encodes a 3,241-amino-acid protein with an overall molecular mass of 335,956 Da. The protein is larger than any bacterial exotoxins identified to date and shows no sequence similarity to other known leukotoxins. Thus, this protein may represent a new class of bacterial leukotoxins. The protein is unusual in that it is devoid of cysteine. This is not a characteristic of proteins from anaerobes, as evidenced by the normal content of cysteine residues in the clostridial toxins, including Clostridium botulinum neurotoxin, Clostridium difficile cytotoxin B, Clostridium septicum alpha toxin, and Clostridium tetani tetanus toxin (GenBank accession no. AB037166, AB217292, D17668, and X06214, respectively). The leukotoxin protein has a sequence at its N terminus that has the properties of a signal sequence. This may indicate that the protein is exported across the cytoplasmic membrane in F. necrophorum in a Sec pathway-dependent manner.

The DNA sequences flanking lktA suggest that this toxin gene may be part of a multigene operon with at least one ORF upstream and another downstream of this gene. The activity of the LktA protein expressed in E. coli indicates that the other proteins encoded in the putative leukotoxin operon are not required to produce a biologically active toxin. Their role may be in secretion of the toxin across the cytoplasmic and outer membranes of F. necrophorum into the culture fluid.

If the lktA determinant is part of an operon, it would be greater than 12 kb in length. A dilemma with such a large operon might be to efficiently translate the mRNA species without premature dissociation of the ribosome from the message. A peculiarity in the cloned region is an abundance of potential RBS sequences. Within the cloned region, there are 26 occurrences of GGAGG, which is a perfect match to the sequence at the 3' end of the 16S rRNA (2, 16). The complementary sequence, CCTCC, which has the same G+C content but which does not act as an RBS, is present only two times in the sequence. The abundance of the GGAGG sequence could provide translation reinforcement sequences to help ensure that a ribosome remains associated with the message and completes the translation of the ORFs. The abundance of the putative RBS sequence (GGAGG) is due to the presence of diglycine repeats in the amino acid sequence. The GGA glycine codon occurs 263 times in the leukotoxin ORF, and 24 of the 26 occurrences of GGAGG in the 11,130 bp sequenced to date correspond to tandem repeats of this codon. The presence of the diglycine repeats in the protein may provide the additional benefit of enabling more-efficient translation of the message.

Expressing the 3.5-kb sequence from the 5' end of lktA caused immediate cessation of growth and lysis of E. coli carrying this recombinant expression vector. Creation of overlapping truncations allowed the expression of the entire leukotoxin gene without significant toxicity to the E. coli host cells. Polyclonal antileukotoxin antiserum reacted strongly to three truncated polypeptides (BSBSE, SX, and FINAL) and more weakly to the other two truncated polypeptides (GAS and SH) in Western blot analysis. This low reactivity was not due to the poor immunogenicity of these relatively hydrophobic polypeptides, because both polypeptides (GAS and SH) produced high antibody titers in rabbits. Thus, it may been due to the tertiary folding pattern of leukotoxin under native conditions. The toxin, a secreted protein, would have its hydrophobic domains internalized when the protein was properly folded. The epitopes corresponding to these domains may not be as accessible to the immune system. Antibodies against these epitopes would thus be underrepresented when the whole undenatured toxin is used as the immunogen. Interestingly, antibodies to one of these polypeptides, GAS, was neutralizing. Thus at least some of the critical epitopes are available in the active toxin.

The intact leukotoxin gene was introduced into E. coli under the control of the lac promoter. Inducible expression of the full-length leukotoxin protein was achieved without any recognizable toxicity to E. coli host cells. Expression of the full-length leukotoxin instead of truncated polypeptides may allow correct folding of the toxin. This would result in internalization of the hydrophobic domains, with a corresponding reduction of toxicity in E. coli host cells. Both polyclonal and monoclonal antibodies against native leukotoxin recognized a protein species with a size consistent with that of the intact leukotoxin in Western blot analysis of cell lysates of E. coli harboring pSN2000. Antibodies raised against all five truncated leukotoxin polypeptides, but not the upstream polypeptide, recognized full-length recombinant leukotoxin as well.

To determine the prevalence and heterogeneity of the leukotoxin gene in this species, 15 strains belonging to F. necrophorum subsp. necrophorum and F. necrophorum subsp. funduliforme isolated from liver abscesses (opportunistic pathogen) or rumen contents (normal inhabitant) were screened for lktA by Southern blotting. Strains belonging to F. necrophorum subsp. necrophorum, irrespective of its location of isolation (liver abscess or ruminal contents), had similar hybridizing patterns. Similarly, all strains of F. necrophorum subsp. funduliforme, irrespective of the site from which it was isolated had identical hybridization patterns, but patterns which differed from the F. necrophorum subsp. necrophorum patterns. The difference in Southern blot hybridization patterns suggests that the disparity in levels of leukotoxin produced between the two subspecies may be due to differences in genetic organization of the leukotoxin locus. Future studies including sequence extension to include the entire leukotoxin operon and identification and characterization of the function of the promoter sequence(s) should help in the understanding of the basis for differential expression of the leukotoxin gene by the two subspecies of F. necrophorum.