The choice of predicted surface-CDS was mostly based on cellular localization since surface proteins are potential targets for mediating adhesion to host. Thus, the LIC10368 CDS was predicted to be an outer membrane protein (93.1%) according to PSORT program 16. The LipoP server predicted LIC10368 CDS to be a lipoprotein with a cleavage site for signal peptidase II at amino acids 18–19 17. It is a hypothetical protein with no known domain by BLAST 1819 and PFAM analysis 20. Identical predicted coding sequence of LIC10368 was identified in L. interrogans serovar Lai 21 but absent in L. borgpetersenii 22 and L. biflexa 23 genome sequences. The gene was amplified, without the signal peptide sequence, and the DNA insert cloned and expressed as a full-length protein in E. coli. Recombinant protein was expressed with 6XHis tag at the N-terminal, purified by metal chelation chromatography, and an aliquot of each step of the process was analyzed through SDS-PAGE (Fig. 1A). The expected protein band of 21 kDa is shown in NaCl-E. coli BL21 (SI)-induced culture, in insoluble form, as inclusion bodies (Fig. 1A, lane 5). Although the Lsa21 protein presented low expression level, it was consistently recovered from the column as single major band (Fig. 1A, lane 6). Structural integrity of the purified protein was assessed by circular dichroism (CD) spectroscopy. As depicted in Fig. 1B, the recombinant protein may encompass a mixture of both α-helices and β-sheets in its secondary structure composition. The experimental data obtained with Lsa21 is similar to the secondary structure content of the native protein predicted by computational analysis 24.

The presence of LIC10368 gene in five pathogenic strains and in one saprophytic strain was examined by PCR with a pair of primers designed according to L. interrogans serovar Copenhageni genome sequences. A 540-bp DNA fragment covering almost the entire LIC10368 predicted CDS was amplified by PCR in all five strains belonging to the pathogenic species L. interrogans serovars: Canicola, Copenhageni, Hardjo, Icterohaemorrhagiae and Pomona. No amplification product was detected in the non-pathogenic L. biflexa serovar Patoc strain Patoc 1 (Fig. 2A). Fragments excised from the gel were purified, cloned into the pGEM-T Easy vector (Promega), and inserts were sequenced. Multiple sequence alignment of the deduced amino acid sequences of the five serovars harboring the LIC10368 gene revealed high conserved identity among the tested strains (not shown).

The ability of both high- (>200) and low- (< 3) passage in vitro cultured leptospires to express LIC10368 was assessed by PCR amplification of reversely transcribed total RNA. Significant levels of gene product could be detected in all three low-passage strains tested (Fig. 2B), whereas among the high-passage strains detectable amounts of LIC10368 transcripts were only observed for the Hond Utrechet IV strain (serovar Canicola). Integrity of total RNA used in RT-PCR experiments was assured by the presence of a 1,042-bp 16S ribosomal cDNA fragment in all samples (Fig. 2B).

It is well known that prolonged culture promotes leptospiral virulence attenuation 25, and it has been suggested that expression of virulence factors may be downregulated upon sequential in vitro culture passages 26. To evaluate the effect of culture-attenuation by in vitro passages on LIC10368 expression, the virulent low-passage strains L. interrogans serovar Pomona strain LPF, L. interrogans serovar Canicola strain LO4, and the reference high-passage strains L. interrogans serovar Pomona strain Pomona, L. interrogans serovar Canicola strain Hond Utrechet IV were grown to late-log phase and collected for RNA isolation. RT-PCR analyses of LIC10368 transcripts were performed, and a significant reduction of gene expression was already observed after 4 (strain LPF) or 3 (strain LO4) passages, thus indicating that LIC10368 expression is rapidly reduced with sequential in vitro culture passages. Amplified products were not detected in control reactions lacking reverse transcriptase, ruling out DNA contamination.

As our results pointed to a correlation between virulence and LIC10368 expression, we decided to examine whether environmental factors, such as osmolarity and temperature could influence LIC10368 regulation at the transcriptional level. All RT-PCR assays were performed with the virulent, low-passage L. interrogans serovar Pomona strain LPF. Induction of LIC10368 expression by osmolarity was assessed by growing cultures at 29°C in EMJH supplemented with 1% rabbit serum and resuspending them in fresh EMJH medium or in EMJH containing 120 mM NaCl. The addition of 120 mM NaCl to the medium mimics physiological conditions (~300 mosmol per liter) encountered by leptospires upon entry into the host 26. Cultures were incubated for 24 h and examined for gene expression. LIC10368 expression was also evaluated in a culture grown under our standard laboratory conditions. As a positive control, we included RT-PCR data for the lipL53 gene (LIC12099) (Fig. 3B). This gene codes for a predicted lipoprotein reactive with sera from patients with leptospirosis 15 and was shown to be strongly up-regulated by sodium chloride in a recently published paper 27. As shown in Fig. 3B, LIC10368 expression appears to be regulated by the osmolarity of the growth medium, because transcripts were detected only upon incubation with additional NaCl, or with a rich EMJH medium containing more serum and salts. Similar up regulation was observed with lipL53 expression, as previously detected by microarrays analysis 27.

Temperature is another important environmental signal that may affect protein expression in bacteria. The transition of temperatures from ambient to mammalian body (37° and later 39°C or higher during the febrile stage) has been correlated with changes in the expression of virulence determinants in many pathogens 28. Moreover, it has been demonstrated that L. interrogans also regulates protein synthesis in response to in vitro temperature changes 2930. Therefore, we compared LIC10368 gene expression patterns of cultures grown at 20°C, 29°C and 37°C, reflecting ambient temperatures in the environment, growth under laboratory conditions, and mammalian host body core temperature. Transcriptional analyses were also performed with cultures subjected to temperature upshifts from 29°C to 37°C and from 37°C to 39°C during an overnight period to simulate conditions experienced by leptospires in the early stages of infection and during febrile stage. LIC10368 transcript levels gradually increased when cultures grown at 20°C, 29°C and 37°C were compared (Fig. 3C). An overnight upshift from 29°C to 37°C had no effect on LIC10368 expression, while an upshift from 37° to 39°C seems to have down-regulated gene expression (Fig. 3C). Our results indicate that temperature is another critical parameter involved in the control of LIC10368 gene expression. Densitometric readings provided information regarding accumulation of LIC10368 transcripts (Fig. 3A, B, C).

Indirect immunofluorescence was used to examine whether Lsa21 protein is surface-exposed. A similar protocol has been previously employed to evaluate the capability of Leptospira to bind factor H 12. The low-passage L. interrogans serovar Pomona strain LPF reacted with antiserum directed against recombinant Lsa21 protein, but no staining was observed with the L. biflexa serovar Patoc strain Patoc 1 (Fig. 4). Pre-immune serum failed to elicit a positive signal in both strains (data not shown).

As Lsa21 protein is suggested to be surface-exposed and might play a role in virulence, we queried whether it could mediate host colonization by adhering to extracellular matrix proteins. Therefore, we examined its interaction with laminin, collagen type I, collagen type IV, cellular fibronectin, and plasma fibronectin. BSA and fetuin were included as negative controls and binding assays were performed by an ELISA-based assay 9. As shown in Fig. 5, Lsa21 protein bound to all immobilized ECM macromolecules tested, but no interaction was detected with BSA or fetuin. As a negative control, we have included another recombinant protein, rLIC10191, known as Loa22 31, and recently described as a virulence factor in Leptospira species 32. A stronger interaction was observed with laminin, collagen IV and plasma fibronectin. This interaction was also assessed on a quantitative basis as illustrated in Fig. 6A. A dose-dependent binding was observed when increasing concentrations of the recombinant protein (0 to 2,000 nM) were allowed to adhere to a fixed laminin, collagen IV or fibronectin concentration (1 μg).

To investigate the role of laminin carbohydrate moieties in the interaction with Lsa21 protein, laminin was oxidized by increasing concentrations of sodium metaperiodate (5 to 100 mM) for 15 min at 4°C. The mild treatment ensures cleavage of vicinal carbohydrate hydroxyl groups, but the polypeptide chain structure remains intact 33. Oxidation effect was dose-dependent and a significant reduction (~65%) in Lsa21 protein attachment to metaperiodate-treated laminin was observed at a 100 mM concentration of periodate (Fig. 6B). These data indicate that laminin sugar residues are critical for the interaction of Lsa21 protein with this major ECM glycoprotein.

To investigate whether Lsa21 protein shared a sequence similarity or a common domain with other previously identified leptospiral ECM- binding proteins 91013 we proceeded with a sequence analysis using Clustal X program and a tree-display NJ plot 3435. The calculated tree derived from sequence alignment is depicted in Fig. 7 and clearly shows that Lsa21 sequence is unrelated to both Len protein family and LigA/LigB proteins. Contrasting with the Len and Lig proteins that divide a common branch inside their family, Lsa21 appears alone on its branch.

IHC (immunohistochemistry) analysis showed essentially similar results in all 5 patients diagnosed with Weil's syndrome. It is necessary to point out that, due to the relatively short interval between autopsy and the patients' death, tissues were better preserved, and classical findings such as liver-plate disarray, although present, were not prominent. Leptospiral Lsa21 antigen was frequently detected as red deposits in the cytoplasm of isolated macrophages along the sinusoidal lining (Kupffer cells and circulating macrophages) (Fig. 7A). Similar deposits were rarely observed in macrophages of the portal inflammatory infiltrate. Lsa21 antigen was also observed focally on the sinusoidal side of the membrane of hepatocytes and as small granules in the underneath cytoplasm (Fig. 7A). Lsa21 protein was infrequently seen over endothelial cells of portal vessels and supra-hepatic branches. In the kidney Lsa21 antigen was present in the luminal side of epithelial cells of the distal nephron, particularly distal tubules and collecting ducts (Fig. 7B, 7C). As observed in liver cells, antigen granules were also identified in the cytoplasm below the cellular membrane. No definite Lsa21 expression was detected in epithelial cells of the proximal tubules. Lsa21 deposit in the cytoplasm of macrophages of the focal interstitial inflammatory infiltrate was rarely observed (see Fig. 8).

The non-pathogenic L. biflexa serovar Patoc strain Patoc 1, the high-passage L. interrogans serovar Canicola reference strains Hond Utrechet IV, L. interrogans serovar Copenhageni strain M-20, L. interrogans serovar Hardjo strain Hardjoprajtino, L. interrogans serovar Icterohaemorrhagiae strain RGA, L. interrogans serovar Pomona, the virulent strains of L. interrogans serovar Pomona strain LPF, L. interrogans serovar Canicola strain LO4 and L. interrogans serovar Copenhageni strain Fiocruz L1-130 were cultured at 29°C under aerobic conditions in liquid EMJH medium (Difco^®- USA) with 10% rabbit serum, enriched with L-asparagine (wt/vol: 0.015%), sodium pyruvate (wt/vol: 0.001%), calcium chloride (wt/vol: 0,001%), magnesium chloride (wt/vol: 0.001%), peptone (wt/vol: 0.03%) and meat extract (wt/vol: 0.02%). Virulence of the L. interrogans serovar Pomona strain LPF, L. interrogans serovar Canicola LO4 and L. interrogans serovar Copenhageni strain Fiocruz L1-130 is maintained by iterative passages in Golden Syrian hamsters. Animals were injected with 10 to 10⁵ leptospires inocula from low and high passage cultures. Hamsters were monitored during 21 days. No significant difference in mortality taxes was observed when animals were inoculated with "no-passage" or "3–4" passages of Leptospira cultures and no death was observed after animal inoculation with high passage cultures. Regulation of LIC10368 expression by environmental factors such as temperature and osmolarity was evaluated in the low-passage L. interrogans serovar Pomona strain LPF. The effect of temperature on LIC10368 transcript levels was assessed by culturing leptospires at 20°C, 29°C and 37°C. Additional cultures grown at 29°C and at 37°C were shifted overnight to 37°C and to 39°C, respectively, to simulate conditions encountered by bacteria upon entry into the host and in a febrile stage. Induction of LIC10368 expression by osmolarity was examined by centrifugation of cultures grown at 29°C in EMJH supplemented with 1% rabbit serum followed by resuspension in fresh EMJH medium or in EMJH containing 120 mM NaCl. Cultures were incubated for 24 h before being harvested for RNA isolation.

Predicted coding sequence (CDS) LIC10368 was selected from the L. interrogans serovar Copenhageni genome sequences based on its cellular localization prediction by PSORT program 16. Sequence motifs, including signal peptides, lipoprotein cleavage sites and transmembrane domains were searched using specific softwares 171819205152. Sequence analysis was performed with Clustal X 35 and a tree displayed program by the Neighbour- Joining method 34.

Predicted CDS LIC10368 was amplified by the PCR from total L. interrogans serovar Copenhageni strain Fiocruz L1-130 genomic DNA using the primer pairs 5'GATGAAAAAAAAGAAAATGAATTGAG and 5'AACGCGATTCATAGAGAGCG. PCR product was cloned into pENTR TOPO vector (Invitrogen) followed by transfer/recombination of DNA insert into the E. coli expression vector pDEST17 using the LR Clonase (Invitrogen). The construct was verified by DNA sequencing with appropriate vector-specific primers. Protein expression was achieved in E. coli BL21 (SI) strain by the action of T7 DNA polymerase under control of the osmotically induced promoter proU 53. The cells were harvested by centrifugation, the bacterial pellets resuspended in 30 ml lysis buffer (500 mM NaCl, 20 mM Tris-HCl, and 0.1% Triton X-100, pH 8.0) and lysed by French Pressure (Aminco). The insoluble fraction was washed 3 times with 30 ml of buffer (20 mM Tris-HCl, 500 mM NaCl, 1 mM β mercaptoethanol, 1 M urea and 1% Triton X-100) before solubilization with 50 ml of buffer containing 20 mM Tris-HCl, 500 mM NaCl, 5 mM β mercaptoethanol, and 6 M guanidine-Cl. Protein refolding was achieved by 500× dilution with 50 mM Tris-HCl (pH 9.0) and 500 mM NaCl. The refolded protein was then purified through metal chelating chromatography in a Sepharose fast flow column (GE Healthcare), extensively dialyzed against phosphate-buffered saline (PBS) pH 7.4, 0.1% (wt/vol) glycine solution for 24 – 48 h and fractions were analyzed in 15% SDS-PAGE.

CD spectroscopy measurements were performed at 20°C in a Jasco J-810 Spectropolarimeter (Japan Spectroscopic, Tokyo, Japan) equipped with a Peltier unit for temperature control. Far-UV CD spectra were measured using a 1 mm path length cell at 0.5 nm intervals. The spectrum is presented as an average of five scans recorded from 190 to 260 nm.

Leptospira cultures were harvested by centrifugation at 11,500 g for 30 min and gently washed in sterile PBS twice. Genomic DNA was isolated from the pellets with a guanidine-detergent lysing solution (DNAzol^® Reagent, Invitrogen), according to manufacturer's instructions. A 540-bp LIC10368 DNA fragment was amplified using oligonucleotides LIC10368-F: 5'GATGAAAAAAAAGAAAATGAATTGAG and R: 5' CTTCGCAACTTGTGGATAAGG and a 1450-bp LIC12099 DNA fragment was amplified using oligonucleotides LP14-F: 5' CACCACCAATGTGTTTGGTATAGCG and LP14-R: 5' CAGCGTTTTGTGATAAAATTAAC designed according to L. interrogans serovar Copenhageni genome sequences (GenBank accession AE016823). PCR was performed in a reaction volume of 25 μl containing 100 ng of genomic DNA, 1 × PCR buffer (20 mM Tris-HCl (pH 8.4), 50 mM KCl), 2 mM MgCl₂, 20 pmol of each specific primer, 200 μM of each dNTP, and 2.5 U Taq DNA Polymerase (Invitrogen). Cycling conditions were: 94°C, 5 min, followed by 35 cycles at 94°C, 50 sec, 60°C (LIC10368)/55°C (LIC12099), 50 sec, 72°C, 1 min 30 sec, and a final extension cycle of 7 min at 72°C. Amplicons were loaded on a 1.5% agarose gel for electrophoresis and visualization with ethidium bromide. Gel-purified bands (Concert Rapid Gel Extraction System – Gibco BRL, Life Technologies) were cloned into the pGEM-T Easy vector (Promega, Madison, Wis.) and the products were sequenced with the primers M13F (5^'-GTTTTCCCAGTCACGA) and M13R (5^'-CAGGAAACAGCTATGAC) on an ABI Prism 3730 × l sequencer (SeqWright, Houston, Tex.). Multiple sequence alignment was performed with Clustal W 54.

Leptospira cultures were harvested by centrifugation at 11,500 g for 30 min. For reverse transcription (RT)-PCR, total RNA was isolated from leptospires cultured as previously mentioned by the acid guanidinium thiocyanate phenol-chloroform method using TRIzol^® Reagent (Invitrogen) according to the manufacturer's recommendations. One microgram of RNA from each sample was treated with 1 U of DNAse I (Invitrogen) for 15 min at RT. DNAse I was inactivated by the addition of 1 μl of 25 mM EDTA solution followed by an incubation at 65°C for 10 min. DNAse-treated RNAs were reversely transcribed using the SuperScript™ III First-Strand Synthesis System for RT-PCR (Invitrogen). One tenth of RT products were amplified in a 25 μl reaction mix using oligonucleotides LIC10368-F/LIC10368-R or LP14-F/LP14-R as described above. In experiments designed to evaluate regulation of LIC10368 expression by environmental factors (osmolarity and temperature), 30 PCR cycles, instead of 35, were employed. Samples quantity and integrity were verified by amplification of a 1,042 bp 16S ribosomal cDNA fragment using oligomers 16S-F 5'CAAGTCAAGCGGAGTAGCAATACTCAGC and 16S-R 5'GATGGCAACATAAGGTGAGGGTTGC. Cycling conditions were: 94°C, 5 min, followed by 30 cycles at 94°C, 50 sec, 62°C, 50 sec, 72°C, 1 min 30 sec, and a final extension cycle of 7 min at 72°C. All samples were tested with and without reverse transcriptase to rule out genomic DNA contamination. PCR-amplified products were loaded onto 1.5% agarose gels for electrophoresis and visualized by ethidium bromide staining. Direct sequencing was performed to confirm the identity of RT-PCR products. Densitometric readings were taken using Image-Quant 2.1 software (Amersham Pharmacia Biotech, Piscataway, NJ). Optical densities of LIC10368 transcripts were normalized for each sample with corresponding 16S ribosomal densitometry data to obtain the relative expression levels.

Ten female BALB/c mice (4–6 weeks old) were immunized subcutaneously with 10 μg of Lsa21 protein. The recombinant protein was adsorbed in 10% (vol/vol) of Alhydrogel (2% Al(OH)₃, Biosector, formerly Superfos Biosector, Denmark), used as adjuvant. Two subsequent booster injections were given at 2-week intervals with the same protein preparation. Negative-control mice were injected with PBS. One week after each immunization, the mice were bled from the retro-orbital plexus and the pooled sera were analyzed by enzyme-linked immunosorbent assay (ELISA) for determination of antibody titers.

The non-pathogenic L. biflexa serovar Patoc strain Patoc 1 and the virulent L. interrogans serovar Pomona strain LPF were grown to late log phase and collected by centrifugation at 11,500 g for 30 min. The cell pellets were washed three times with phosphate-buffered saline (PBS) (pH 7.4) and resuspended in a fixing solution containing 25% glutaraldehyde, 10% formalin, and PBS (pH 7.4), essentially according to the methodology described in Verma et al. 12. Ten microliters of 1 × 10⁹ leptospire per ml suspensions were applied to glass slides, which were then placed on ice for 1 h. After two washes with PBS, blocking was performed using 2% (wt/vol) BSA in PBS for 1 h. Slides were washed twice with PBS and incubated with a 1:50 dilution of mouse anti-Lsa21 protein serum or with preimmune serum for 1 h at room temperature (RT) in a humidifying chamber. After three washes with PBS, slides were incubated with a 1:200 dilution of a fluorescein isothiocyanate-conjugated goat anti-mouse antibody (Sigma) for 1 h at RT. Slides were washed three times with PBS and mounting was performed in the antifading agent Mowiol (Calbiochem) with 2.5% DABCO (Sigma). A coverslip was added, and the slide was left overnight at 4°C before visualization by confocal microscopy (Carl Zeiss, Inc., Jena, Germany). Images were obtained with LSM 510 Meta software and the objective used was C-Apochromat 63×, scan zoom 2.0.

All macromolecules, including the control proteins fetuin and BSA, were purchased from Sigma Chemical Co. (St. Louis, Mo.). Laminin-1 and collagen Type IV were from the basement membrane of Engelbreth-Holm-Swarm mouse sarcoma, cellular fibronectin was from human foreskin fibroblasts, plasma fibronectin was from human plasma and collagen Type I was from a rat tail. Protein attachment to individual macromolecules of the extracellular matrix was analyzed according to a previously published protocol Briefly, ELISA plate wells (Nunc-Immuno Plate MaxiSorp Surface) were coated with 1 μg of laminin, collagen Type I, collagen Type IV, cellular fibronectin, plasma fibronectin, BSA (nonglycosylated attachment-negative control protein) and fetuin (highly glycosylated attachment-negative control protein) in 100 μl of PBS for 2 h at 37°C. The wells were washed three times with PBS-0.05% Tween 20 (PBST) and then blocked with 200 μl of 1% BSA for 1 h at 37°C followed by an overnight incubation at 4°C. One microgram of recombinant protein was added per well in 100 μl of PBS and protein was allowed to attach to the different substrates for 1 h 30 min at 37°C. After washing six times with PBST, bound protein was detected by adding 100 μl of a 1:5,000 dilution of mouse anti-Lsa21 serum in PBS. Incubation proceeded for 1 h and after three washes with PBST, 100 μl of a 1:5,000 dilution of horseradish peroxidase (HRP)-conjugated goat anti-mouse immunoglobulin G (IgG) in PBS were added per well for 1 h. All incubations took place at 37°C. The wells were washed three times, and o-phenylenediamine (0.04%) in citrate phosphate buffer (pH 5.0) plus 0.01% H₂O₂ was added. The reaction was allowed to proceed for 15 min and was then interrupted by the addition of 50 μl of 8 M H₂SO₄. The absorbance at 492 nm was determined in a microplate reader (Labsystems Uniscience, Multiskan EX). For determination of dose-dependent attachment of Lsa21 protein to laminin, collagen IV and plasma fibronectin, protein concentrations varying from 0 to 2,000 nM in PBS were used.

For statistical analyses, the attachment of Lsa21 protein to ECM macromolecules was compared to its binding to BSA by the Student's two-tailed t -test. A P value less than 0.01 was considered statistically significant.

Laminin oxidation was performed as described elsewhere 9. Briefly, microtitre wells were coated with 1 μg of laminin in 50 mM sodium acetate buffer, pH 5.0, and incubated for 16 h at 4°C. Wells were washed three times with 50 mM sodium acetate buffer, pH 5.0, and immobilized laminin was treated with different sodium metaperiodate concentrations (5 – 100 mM) in the same buffer for 15 min at 4°C in the dark. After three washes with 50 mM sodium acetate buffer, wells were blocked with 200 μl of 1% BSA for 1 h at 37°C. Binding of Lsa21 protein (1 μg in PBS per well) to periodate-treated laminin was assessed as outlined above.

Five patients, three males and two females, were autopsied with a diagnosis of Weil's syndrome. The age of the four patients ranged from 20 to 29 years old and one patient was 67 years old. Illness duration averaged seven days with a sudden onset manifested by headache, malaise, muscle pains and high fever. They developed jaundice, palpable enlarged liver, acute renal failure and a hemorrhagic syndrome. Epidemiological data showed that patients had had contact with contaminated water, mainly from floods and sewage. Main clinico-epidemiological data of the patients are in Table 1. Of note, two cases (4 and 5) were part of our previous publication 47. The illness was usually of short duration and this, associated with delayed clinical diagnosis, contributed to the lack of some important confirmatory laboratorial tests. In fatal human cases and in farm animals, diagnosis by immunohistochemistry methods are specific and also offer clues to the understanding of the pathogenesis of the disease 47. Special care was taken in selecting autopsies with typical macro and microscopical findings described in leptospirosis and performed after an average of up to five to six hours of death. Liver and kidney samples, which are more representative of the disease, were routinely fixed in formalin, embedded in paraffin and stained hematoxylin-eosin. Liver and kidney of five non- leptospirotic patients were obtained from autopsies performed within a close post-mortem interval as compared to leptospirotic patients and used as controls. Leptospiral antigens were detected by an immunohistochemical assay using polyclonal rabbit anti-serum as previously described 47 with positive results in all five leptospirotic patients. Controls were negative.

IHC studies were performed to assess the presence and localization of leptospiral antigens as detected by the mouse anti-Lsa21 serum (1:1,000 dilution). Kidney and liver 3 μm sections were analyzed using a standard immunohistochemistry protocol 47 with EnVision labeled polymer-AP mouse/rabbit K 4018 55. Antigen retrieval was performed by steamer incubation of the sections in citrate pH 6.0. Staining was completed with liquid permanent Red Chromogen K0640 55. All specimens were then lightly counterstained with hematoxylin. This method represents a better choice for diagnosis than the usual silver stains to detect the microorganisms particularly when we are dealing with human autopsies.

All animal studies were approved by the Ethics Committee of the Instituto Butantan (Sao Paulo, Brazil). Autopsies were performed at the Department of Pathology, University of Sao Paulo, School of Medicine (Sao Paulo, Brazil) and are routinely used for diagnosis, teaching and research purposes. All protocols used were verbally approved by the Ethics Committee of the University of Sao Paulo (São Paulo, Brazil).