Rickettsioses in South Korea, Data Analysis

To the Editor: Choi et al. (1) conducted a study on sequence analysis of a partial rompB gene amplified from sera of humans who were seropositive for spotted fever group (SFG) and typhus group rickettsioses. They write, "These finding suggested that several kinds of rickettsial diseases, including boutonneuse fever, rickettsialpox, R. felis infection, and Japanese spotted fever… are occurring in Korea." 
 
These claims propagate some errors and may lead to an inadequate conclusion. First, rompB is conserved in Rickettsia spp. and consists of 4,968 bp with respect to the published sequence of the R. conorii strain Seven (2,3). Fournier et al. (4) amplified 4,682 bp of rompB and showed a high degree of nucleotide sequence similarity (99.2%) between R. africae and R. sibirica, R. pakeri, and R. slovaca. Choi et al. amplified ≈420 bp of rompB (position 3562–4077) for sequence analysis. This segment is located in a highly conserved region of the gene, which may decrease the ability to differentiate particular species from other SFG rickettsiae. This study cannot prove the existence of specific SFG rickettsioses until the results are confirmed further by, for example, isolating these SFG rickettsiae from humans, animals, or ticks in South Korea. Recently, the authors analyzed nucleotide sequences of 267-bp amplicons of rompB (position 4762–4496) obtained from patient sera and found that R. conorii could not be differentiated from R. sibirica (5). This finding also supports our concerns. 
 
Second, although partial rompB nucleotide sequence analysis of rickettsiae obtained from 1 patient's serum showed 98.87% similarity with R. conorii strain Seven, the finding does not indicate boutonneuse fever is occurring in South Korea because high similarities (98.6%–99.8%) are found among 4 subspecies of R. conorii. Multilocus sequence typing can help differentiate among these subspecies (6). 
 
This study provided a model to amplify SFG rickettsial DNA from sera of patients, and it will be helpful in surveillance of these diseases. However, the results should be interpreted more carefully in the context of clinical and epidemiologic data and combined with different gene sequence analyses to obtain a reliable and specific diagnosis.

Rickettsia species they detected. Therefore, apart from R. felis, which was not used as a positive control, positive products may result from cross-contamination. Finally, technically, the data are impossible to reproduce: 1) primer sets WJ77/80 and WJ79/83/78 cited in the legends of Figures 2 and 3 are neither described nor referenced in the text, 2) sequence of the RpCS.877p primer in Table 1 differs from that in the referenced article (4), 3) described sequences have not been deposited in GenBank, and 4) all rompB primers described in Table 1 exhibit 1-6 nucleotide mismatches with ompB sequences of at least 1 of the detected species. Based on these errors, the 7 cases of dual infections with R. conorii and R. typhi, which have never been reported before, are doubtful, and these data need to be confirmed.

Rickettsioses in South Korea, Data Analysis
To the Editor: Choi et al.
(1) conducted a study on sequence analysis of a partial rompB gene amplified from sera of humans who were seropositive for spotted fever group (SFG) and typhus group rickettsioses. They write, "These finding suggested that several kinds of rickettsial diseases, including boutonneuse fever, rickettsialpox, R. felis infection, and Japanese spotted fever… are occurring in Korea." These claims propagate some errors and may lead to an inadequate conclusion. First, rompB is conserved in Rickettsia spp. and consists of 4,968 bp with respect to the published sequence of the R. conorii strain Seven (2,3). Fournier et al. (4) amplified 4,682 bp of rompB and showed a high degree of nucleotide sequence similarity (99.2%) between R. africae and R. sibirica, R. pakeri, and R. slovaca. Choi et al. amplified ≈420 bp of rompB (position 3562-4077) for sequence analysis. This segment is located in a highly conserved region of the gene, which may decrease the ability to differentiate particular species from other SFG rickettsiae. This study cannot prove the existence of specific SFG rickettsioses until the results are confirmed further by, for example, isolating these SFG rickettsiae from humans, animals, or ticks in South Korea. Recently, the authors analyzed nucleotide sequences of 267-bp amplicons of rompB (position 4762-4496) obtained from patient sera and found that R. conorii could not be differentiated from R. sibirica (5). This finding also supports our concerns.
Second, although partial rompB nucleotide sequence analysis of rickettsiae obtained from 1 patient's serum showed 98.87% similarity with R. conorii strain Seven, the finding does not indicate boutonneuse fever is occurring in South Korea because high similarities (98.6%-99.8%) are found among 4 subspecies of R. conorii. Multilocus sequence typing can help differentiate among these subspecies (6).
This study provided a model to amplify SFG rickettsial DNA from sera of patients, and it will be helpful in surveillance of these diseases. However, the results should be interpreted more carefully in the context of clinical and epidemiologic data and combined with different gene sequence analyses to obtain a reliable and specific diagnosis.

"Mycobacterium tilburgii " Infections
To the Editor: Advanced molecular biologic methods have improved the species differentiation and taxonomic classification of microorganisms, including nontuberculous mycobacteria. Identifying and characterizing an increasing number of "new" mycobacterial species of medical importance is now possible. Often, these newly described mycobacteria have been isolated from immunocompromised patients (1,2). Some of those have been difficult to cultivate, and 16S rRNA gene sequencing or similar methods have become of major importance to allow species identification and clinical diagnosis. Here, we report 2 patients with disease likely caused by a novel mycobacterial species that could not been previously cultivated. Diagnosis relied on molecular identification of acid-fast organisms in tissues. We also briefly review 2 similar cases published previously and note that all 4 known patients were from central Europe.
A 43-year-old woman without evidence of immunodeficiency reported recurrent episodes of dysuria, hematuria, and abdominal discomfort for >1 year. On cystoscopy, a hyperemic bladder with yellow plaques was observed. Biopsy of the plaques showed granulomatous infiltration of histiocytes. No definite diagnosis was made, and symptomatic relief occurred after a trial of empiric antimicrobial drug therapy. When the patient sought treatment again with persistent abdominal discomfort, endoscopy showed a lesion in the stomach that resembled a healed ulcer and numerous elevated yellow plaques throughout the colon and ileum. Microscopically, a granulomatous inflammation with macrophages filled with many acid-fast rods was seen, but mycobacterial growth did not occur in different media or in a guinea pig. Antituberculous treatment was initiated, and the patient slowly improved. A repeat colonoscopy showed fewer and smaller lesions. Efforts to culture the organism from biopsy specimens were again unsuccessful (different solid and liquid media, blood or chocolate agar, guinea pig, Balb/c mice). Sequencing of polymerase chain reaction (PCR) products of the 16S rRNA gene (3) from the organism represented a previously unknown mycobacterial species (EMBL DNA database: accession number Z50172). "Mycobacterium tilburgii" was proposed as the designation of this species because the novel mycobacterium was identified in the city of Tilburg (4). Retrospective analysis of the initial bladder biopsy specimen and of 2 lymph nodes taken during abdominal surgery (which became necessary because of a complicating ileal obstruction) confirmed the presence of "M. tilburgii" 16S rRNA gene sequences. All samples yielding PCR fragments hybridized with a "M. tilburgii"-specific biotinylated DNA probe.
A 34-year-old AIDS patient sought treatment for involuntary weight loss. Endoscopy showed white superficial embossments of the duodenal mucosa. Biopsy specimens were negative for acid-fast bacteria and mycobacterial growth (again, different solid and liquid media, extended incubation periods). Antiretroviral therapy was begun, but asthenia, persistent fever, diarrhea, vomiting, and cachexia developed. Repeat endoscopy showed yellow, plaquelike lesions in the duodenum (online Appendix Figure; available at http://www.cdc.gov/ncidod/EID/vol1 2no03/05-1139-G1.htm) and esophagus with periodic acid-Schiff (PAS) and acid-fast intracellular bacteria that were nonculturable. PCR of 16S rRNA gene (3) confirmed the presence of mycobacterial DNA; sequencing showed 100% homology to the