Leptospira noguchii and Human and Animal Leptospirosis, Southern Brazil

To the Editor: Pathogenic leptospires, the causative agents of leptospirosis, exhibit wide phenotypic and genotypic variations. They are currently classified into 17 species and >200 serovars (1,2). Most reported cases of leptospirosis in Brazil are of urban origin and caused by Leptospira interrogans (3). Brazil underwent a dramatic demographic transformation due to uncontrolled growth of urban centers during the last 60 years. Urban slums are sites of poor sanitation that favors rat-borne transmission of leptospirosis among humans. Thus, this may explain the major involvement of serovar Copenhageni (L. interrogans). The predominance of L. interrogans is likely due to the underestimation of rural cases of leptospirosis.

previously isolated from animals such as armadillo, toad, spiny rat, opossum, nutria, the least weasel (Mustela nivalis), cattle, and the oriental fi re-bellied toad (Bombina orientalis) in Argentina, Peru, Panama, Barbados, Nicaragua, and the United States (1,6). Human leptospirosis associated with L. noguchii has been reported only in the United States, Peru, and Panama, with the isolation of strains Autumnalis Fort Bragg, Tarassovi Bac 1376, and Undesignated 2050, respectively (1,6). The Fort Bragg strain was isolated during an outbreak among troops at Fort Bragg, North Carolina. It was identifi ed as the causative agent of an illness characterized by fever, headache, myalgia, and a pretibial rash-Fort Bragg fever (7). We were not able to obtain data regarding the other 2 human isolates.
We report the isolation of 3 additional L. noguchii strains from Brazil, including 2 from cases of human leptospirosis. The fi rst isolate (Bonito strain) was obtained from the blood culture of a 34-year-old man who exhibited fever, headache, myalgia, hemorrhages, jaundice, abdominal pain, diarrhea, and vomiting. The patient reported contact with rats, farm animals, and dogs before the onset of illness. Laboratory tests at admission to the Hospital Santa Casa de Misericórdia, Pelotas, showed an elevated level of serum bilirubin (total 21 mg/ dL, direct 16 mg/dL) and a slight increase in liver enzyme levels (alanine aminotransferase 2×, aspartate aminotransferase 1.5× above reference levels). An acute-phase serum sample showed a titer of 25 against serovars Autumnalis and Bratislava by microscopic agglutination test (MAT).
The second isolate (Cascata strain) was obtained from the blood culture of a 16-year-old boy who exhibited headache, fever, fl ulike symptoms, and myalgia. He reported previous contact with rats and dogs. The patient was not hospitalized, and an acute-phase serum sample showed a titer of 25 against saprophytic serovar Andamana by MAT. Both patients were from the rural area of Pelotas. Unfortunately, convalescent-phase serum samples were not obtained from these patients.
A third isolate (Hook strain) was obtained from a male stray dog with anorexia, lethargy, weight loss, disorientation, diarrhea, and vomiting. The animal died as a consequence of the disease. The isolate was obtained from a kidney tissue culture. No temporal or spatial relationship was found between the 3 cases.
Serogrouping was performed by using a panel of rabbit antisera. Bonito, Cascata, and Hook strains were classifi ed as Autumnalis, Bataviae, and Australis, respectively. Serogroups were confi rmed by the strong and specifi c reaction of hyperimmune sera against these isolates, with the reference strains of the respective serogroups. Species identifi cation was accomplished by sequencing nearly the full length of the 16S rRNA gene, as previously described (5). The sequences of the Hook, Cascata, and Bonito strains were deposited in GenBank under accession nos. EU349494-EU349496.
In addition, the rpoB gene sequence was determined and used for further confi rmation of the species. The rpoB sequence for the strains Hook, Cascata, Bonito, and the L. noguchii reference strains were deposited in GenBank under accession nos. EU349497-EU349505. BLAST (www.ncbi.nlm.nih.gov/blast/Blast. cgi) alignment confi rmed the new isolates as L. noguchii. The 16S rRNA gene sequence was also used for taxonomic analysis of L. noguchii (Figure). The topology-based dendrogram demonstrates sequence relatedness among strains isolated in Pelotas and the L. noguchii Autumnalis, Australis, and Bataviae strains deposited in Gen-Bank ( Figure). No molecular or serologic characterization at the serovar level was performed.

Aquaculture and Florfenicol Resistance in Salmonella enterica Serovar Typhimurium DT104
To the Editor: In a letter recently published in Emerging Infectious Diseases, Smith (1) discussed evidence that he mistakenly believes to undermine the hypothesis that the fl orfenicol resistance gene present in some isolates of the epidemic Salmonella enterica serovar Typhimurium DT104 strain originated from a fl orfenicol resistance plasmid present in Vibrio damsela (Pasteurella piscicida) that infected fi sh farms in Japan in the 1990s (2). Smith correctly states that the fl orfenicol resistance gene was present in S. enterica serovar Typhimurium DT104 strains isolated in the United States in 1985, before the gene was documented in V. damsela in Japan (1,3). He is also correct in noting that this particular fl orfenicol resistance gene was detected in a plasmid in Klebsiella pneumoniae in France in  1969 (1,4).
However, an earlier report by Briggs and Fratamico (5) clearly established that the fl orfenicol resistance genes and the tetracycline resistance genes tetG and tetR in the Salmonella genomic island 1 (SGI1) were surrounded by non-antimicrobial-drug resistance DNA. This DNA is homologous to DNA sequences in plasmids PASPPFLO and pJA8122 (see Figure  1 and Table 2 in reference 5) (5-7). In addition to antimicrobial drug resistance genes, PASPPFLO and pJA8122 contain cloned DNA segments of indigenous R plasmids found in V. damsela and V. anguillarum, respectively; these cloned DNA segments span sequences that extend beyond their fl orfenicol resistance and tetR/tetG genes (5-7). For example, the region of the fl orfenicol resistance gene in SGI1 contains 763 nt of the non-antimicrobial-drug resistance portion of the original V. damsela plasmid; the region of tetR/tetG contains 468 nt of the non-antimicrobial-drug resistance DNA segment of the P. piscicida plasmid (5)(6)(7).
The presence of these non-antimicrobial-drug resistance R plasmid DNA sequences in SGI1 constitutes a molecular signature that fi rmly establishes the aquaculture origin of the fl orfenicol resistance and the tetR/ tetG genes in the S. enterica serovar Typhimurium DT104 strain studied by Briggs and Fratamico and in the SGI1 of other bacteria (5). These R plasmid DNA sequences in SGI1 also confi rm direct or indirect horizontal gene transfer between bacteria in the aquaculture environment and S. enterica serovar Typhimurium DT104 (5-7). In Response: In his letter (1), Cabello makes 2 observations regarding the debate concerning the origin of the fl oR gene in Salmonella enterica serovar Typhimurium DT104. The fi rst observation is that the plasmid PASPP-FLO contained cloned segments of an indigenous Vibrio damsela plasmid. However, PASPPFLO is not the name of a plasmid but is the GenBank locus identifi er associated with the sequence (GenBank accession no. D37826) of a 3,745-bp region of the V. damsela plasmid pSP92088 that contained ppfl o (2,3).

Felipe C. Cabello
The second observation is that sequences fl anking the fl oR gene in S. enterica serovar Typhimurium DT104 (GenBank accession no. AF071555) are homologous to those fl anking the pp-fl o gene sequenced from the V. damsela plasmid pSP92088 (4). On the basis of this homology, he seems