Infection by Ralstonia Species in Cystic Fibrosis Patients: Identification of R. pickettii and R. mannitolilytica by Polymerase Chain Reaction

The frequency of respiratory tract infections caused by Ralstonia species in persons with cystic fibrosis (CF) and the role of these species in CF pulmonary disease are not well documented. In part, this lack of documentation may be attributed to the difficulty in accurately identifying Ralstonia species; R. mannitolilytica and R. pickettii in particular may be misidentified as other closely related species, particularly those of the Burkholderia cepacia complex. We used polyphasic analyses to identify 42 Ralstonia isolates from sputum cultures from 38 CF patients. Several isolates that could not be identified to the species level may belong to novel Ralstonia species. We demonstrated chronic colonization by using genotyping of serial isolates recovered from the same patient. To facilitate identification of R. mannitolilytica and R. pickettii, we developed 16S ribosomal DNA-based polymerase chain reaction assays that allow sensitive and specific identification of these species.

The frequency of respiratory tract infections caused by Ralstonia species in persons with cystic fibrosis (CF) and the role of these species in CF pulmonary disease are not well documented. In part, this lack of documentation may be attributed to the difficulty in accurately identifying Ralstonia species; R. mannitolilytica and R. pickettii in particular may be misidentified as other closely related species, particularly those of the Burkholderia cepacia complex. We used polyphasic analyses to identify 42 Ralstonia isolates from sputum cultures from 38 CF patients. Several isolates that could not be identified to the species level may belong to novel Ralstonia species. We demonstrated chronic colonization by using genotyping of serial isolates recovered from the same patient. To facilitate identification of R. mannitolilytica and R. pickettii, we developed 16S ribosomal DNA-based polymerase chain reaction assays that allow sensitive and specific identification of these species.
ystic fibrosis (CF) is the most frequent hereditary disease in Caucasian populations (1); chronic microbial colonization of the large airways, leading to exacerbations of pulmonary infection, is the major cause of illness and death in CF patients. Typical CF pathogens include Staphylococcus aureus, Pseudomonas aeruginosa, Haemophilus influenzae, and Burkholderia cepacia complex; other species, including Stenotrophomonas maltophilia, Alcaligenes (Achromobacter) xylosoxidans, B. gladioli, and R. pickettii have been recovered from sputum cultures of CF patients as well (2,3). Recently, we showed that a number of unusual bacterial species (including several novel species within the α-Proteobacteria) are also occasionally isolated from CF patients (4). Infection with mucoid P. aeruginosa and members of the B. cepacia complex is associated with increased illness and death in CF patients (5-7), but the clinical importance of infection with these other species is less clear.
We describe the occurrence of several Ralstonia species in the respiratory secretions of CF patients. We also describe the development and evaluation of two polymerase chain reaction (PCR) assays for rapid, accurate identification of R. pickettii and R. mannitolilytica.

Bacterial Strains and Study Population
Since early 1997, the Burkholderia cepacia Research Laboratory and Repository (University of Michigan, Ann Arbor, MI) has received more than 4,000 bacterial isolates, collected from CF patients receiving care in 145 CF treatment centers in 130 U. S. cities. Isolates received were tentatively identified by the referring microbiology laboratory as B. cepacia complex or a related species or were not identified to the species level. From these isolates, we identified 42 Ralstonia isolates obtained from 38 patients who had received care in 19 treatment centers in 18 U. S. cities. The type and reference strains of Ralstonia, Pandoraea, Burkholderia, Alcaligenes, and Bordetella species have been described (9)(10)(11)(12)(13)(14). These strains were obtained from the BCCM/LMG-Bacteria Collection RESEARCH (Laboratorium voor Microbiologie, Universiteit Gent, Belgium) or were provided by D. Henry (University of British Columbia, Vancouver, Canada). All isolates were grown aerobically on Mueller-Hinton broth (Becton, Dickinson and Company, Cockeysville, MD) supplemented with 1.8% (wt/vol) agar and incubated at 32°C.

Species Identification
We used a polyphasic approach to identify all isolates, including biochemical tests, 16S ribosomal (r)DNA-based PCR assays and sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) of whole-cell proteins. Biochemical tests (determination of oxidase, lysine decarboxylase, and o-nitrophenyl-β-D-galactoside activity; growth on B. cepacia selective agar; and oxidation-fermentation of sucrose) were performed as described (27). SDS-PAGE of whole-cell proteins was performed as described (9,10), and isolates were identified by comparison with a database containing protein profiles of all Ralstonia species. We used 16S rDNA-based PCR assays (28) to determine whether or not isolates belonged to the genera Burkholderia or Ralstonia or to the B. cepacia complex.

Genotyping of Serial Isolate
Multiple isolates from a single patient were genotyped by randomly amplified polymorphic DNA (RAPD) genotyping as described (29). We digitized gel images with a GelDoc2000 gel analyzer (Bio-Rad Laboratories, Hercules, CA) and stored them as tagged image files. After normalization with the molecular weight marker, patterns were analyzed with Molecular Analyst Fingerprinting Plus software (Bio-Rad Laboratories). Similarities between patterns were calculated by using Pearson's product-moment correlation coefficient. We considered isolates to belong to the same genotype if they shared 90% or more similarity.

Evaluation of the PCR Assays
For evaluating PCR assays, we tested 152 isolates, including 79 Ralstonia isolates (both clinical isolates and reference strains) and 73 isolates representing phylogenetically related species that may be found in sputum cultures of CF patients. Isolates tested were as follows: R. pickettii (27 isolates

Species Identification
Isolates were tentatively identified as belonging to the genus Ralstonia if they 1) reacted with primer pair RHG-F/ RHG-R (specific for Burkholderia and Ralstonia spp.) (28), 2) showed no lysine decarboxylase and o-nitrophenyl-β-D-galactoside activity, 3) produced no acid from sucrose, and 4) showed oxidase activity. Using these criteria, we identified 42 putative Ralstonia sp. isolates. These isolates were further identified to the species level by using SDS-PAGE of wholecell proteins. Most isolates (25) were identified as R. mannitolilytica; 9 were identified as R. pickettii. Two isolates were identified as R. gilardii, and another as R. taiwanensis. Five isolates clearly belonged to the genus Ralstonia but could not be classified into one of the known species. Pending further investigations, these isolates were classified as Ralstonia sp.

Genotyping of Serial Isolates
We identified two patients (A and B) who were sputumculture positive for R. mannitolilytica and one patient (C) who was culture positive for R. pickettii on more than one occasion. The three R. mannitolilytica isolates cultured from patient A were recovered over a period of >2 years. RAPD genotyping indicated that the first isolate clearly differed from the two isolates recovered subsequently; the latter two isolates (recovered 20 months apart) were the same genotype ( Figure 1). Similarly, the two R. mannitolilytica isolates recovered from patient B (cultured 8 weeks apart) were the same genotype, as were the two R. pickettii isolates recovered from patient C (cultured 6 weeks apart) (Figure 1).

Primer Design
Alignment of 16S rRNA gene sequences of Ralstonia sp. available in GenBank showed similarity values >93.1% and >98.2% within the species R. pickettii and R. mannitolilytica, respectively. Identity of sequences between these two species ranged from 89.9% to 96.8%. Several species-level sequence signatures were detected and were incorporated into the species-specific primers Rp-F1 and Rp-R1 (forward and reverse primer for R. pickettii) and Rm-F1 and Rm-R1 (forward and reverse primer for R. mannitolilytica). PCR with these primers resulted in the amplification of fragments of 210 bp and 398 bp, respectively ( Figure 2). Each of the 152 strains included in this study was examined by PCR with the primer pairs described (Table).

Discussion
The occurrence and clinical role of Ralstonia sp. in the respiratory secretions of persons with CF have not been systematically investigated because of the rapidly changing taxonomy of the genus Ralstonia and the absence of rapid, reliable methods for species identification. We used a polyphasic approach to identify Ralstonia sp. in sputum cultures of CF patients and developed two PCR assays for identifying R. pickettii and R. mannitolilytica.
Previous reports describing the bacterial flora of the respiratory tract of CF patients have focused mainly on P. aeruginosa and B. cepacia complex organisms (3,5,31); reports describing the presence of Ralstonia species in sputum cul-tures of CF patients are scarce and often anecdotal. In a prospective study, Burns et al. (2) isolated R. pickettii from only 2 of 559 patients. More recently, we have shown that other Ralstonia species, including R. mannitolilytica, R. taiwanensis, and R. gilardii, can also be isolated from the respiratory secretions of CF patients (4). In this study, we identified Ralstonia species recovered from sputum cultures of 38 CF patients. Collectively, these data indicate that the prevalence of Ralstonia sp. in the CF population is rather low. However, because we did not specifically survey all referring laboratories for all Ralstonia species that may have been recovered from CF specimens, we were not able to define a more precise prevalence of Ralstonia sp. in the CF population.
Our data do not provide evidence for patient-to-patient spread of Ralstonia sp. because no clustering of cases occurred within centers or geographic regions (data not shown). However, we were able to document persistent colonization with Ralstonia species in three patients. Patient A's infection is particularly interesting. In this patient, an initial R. mannitolilytica strain was apparently replaced with another strain, which then persisted for >20 months. However, the bacterial and host factors involved in infection by more than one R. mannitolilytica strain or with chronic colonization remain to be defined.
Five Ralstonia isolates could not be identified to the species level. 16S rDNA PCR and SDS-PAGE of whole-cell proteins clearly indicated that these isolates belong to the genus Ralstonia, suggesting that they may represent novel Ralstonia sp. Further polyphasic taxonomic studies are needed to clarify their status. The finding of R. mannitolilytica, R. gilardii, R. taiwanensis, and possible novel Ralstonia species in respiratory secretions of CF patients suggests that these organisms may be emerging human pathogens and again highlights the fact that the bacterial biodiversity in the respiratory tract of CF patients has thus far been underestimated (4).
Of the 25 R. mannitolilytica strains identified in this study, 9 were initially identified by the referring laboratory as R. pickettii, 8 as B. cepacia complex, 6 as Burkholderia sp., 1 as B. gladioli, and 1 as P. fluorescens. Of the 9 R. pickettii strains identified, 3 were identified by the referring laboratory as R. pickettii, 2 as Burkholderia sp., 1 as Pseudomonas sp., 1 as B. cepacia complex, and 2 isolates as unidentified. The R. gilardii and R. taiwanensis isolates were received as B. cepacia complex and S. maltophilia, respectively. Most (81%) of these isolates were capable of growth on B. cepacia selective agar. These observations reiterate that identification of these species is not straightforward and that their misidentification as other CF pathogens, such as B. cepacia complex, is not uncommon. Such misidentification has an important impact on infection control in CF since the efficiency of these measures depends on accurate identification of the microorganisms involved. Infection-control policies, particularly those recommended to prevent interpatient spread of B. cepacia complex, have a tremendous impact on the quality of life of CF patients (6,7). To enhance accurate identification of CF pathogens, several PCR assays have been developed recently (28,30,(32)(33)(34)(35). We RESEARCH sought to design similar PCR tests to allow the identification of R. pickettii and R. mannitolilytica based on species-level signature sequences in the 16S rRNA gene. By comparing available R. pickettii and R. mannitolilytica 16S rRNA gene sequences with sequences from other Ralstonia species and representatives of the phylogenetically closely related genera Burkholderia and Pandoraea, we identified several regions that showed sufficient diversity to allow the design of primer pairs Rp-F1/Rp-R1 and Rm-F1/Rm-R1, permitting the sensitive and specific identification of R. pickettii and R. mannitolilytica, respectively (Table).
The results of our study indicate that a number of Ralstonia species can be isolated from sputum cultures of CF patients. The correct identification of these species presents a challenge for diagnostic microbiology laboratories. Our study supports the use of genotypic methods to augment routine phenotypic evaluation. The combined use of the two PCR assays described will allow the identification of most Ralstonia species encountered in sputum cultures of CF patients. Most importantly, the use of these assays will substantially reduce the misidentification of R. pickettii and R. mannitolilytica as B. cepacia complex. These tests will be a valuable adjunct in the evaluation of CF sputum culture isolates and will allow more precise study of the prevalence and natural history of human infection by these emerging pathogens.  All others (n=118) 1 c 117 a Of the three R. pickettii isolates that gave a false-negative reaction in this PCR test, two were identified as R. pickettii by the referring laboratory, while one was received unidentified. b One unidentified Ralstonia sp. isolate cross-reacted with this primer pair. c One R. pickettii isolate cross-reacted with this primer pair.