Journal of Cystic Fibrosis 12 (2013) 298–301 www.elsevier.com/locate/jcf Short Communication Identification and distribution of Achromobacter species in cystic fibrosis ⁎ Theodore Spilker a, Peter Vandamme b, John J. LiPuma a, a Department of Pediatrics and Communicable Diseases, University of Michigan Medical School, Ann Arbor, MI, USA b Laboratory of Microbiology, Ghent University, Ghent, Belgium Received 20 August 2012; received in revised form 3 October 2012; accepted 6 October 2012 Available online 7 November 2012 Abstract Background: We recently described a multilocus sequence typing scheme for Achromobacter that identified several novel species in this genus. Methods: We assessed the ability of nrdA sequence analysis to differentiate Achromobacter species, including the seven previously named species and 14 recently described genogroups. Confirmation of distinctness between groups was confirmed using the k parameter. Using this single locus sequence to differentiate species, we analyzed Achromobacter isolates obtained from 341 CF patients in the U.S. Results: We found that Achromobacter xylosoxidans accounts for 42% of Achromobacter infections, while Achromobacter ruhlandii accounted for 23.5% of infections. Isolates from 17% of patients were members of the novel genogroup 14. The remaining 17.5% of strains belonged to 11 other species/genogroups. Conclusion: The use of nrdA sequence analysis allows differentiation of the several Achromobacter species that can infect persons with CF. Achromobacter species other than A. xylosoxidans account for the majority of Achromobacter infection in CF patients in the U.S. © 2012 European Cystic Fibrosis Society. Published by Elsevier B.V. All rights reserved. Keywords: Achromobacter; Infection; Taxonomy; Clinical microbiology 1. Introduction noted that sequence analysis of a single locus in the MLST scheme, nrdA, appeared to be able to distinguish these 21 Achromobacter xylosoxidans is well recognized as an taxa. We further noted that approximately half of the opportunistic pathogen infecting persons with cystic fibrosis Achromobacter isolates included in the study, the great (CF) [1]. The remaining six currently named Achromobacter majority of which had been recovered from CF and non-CF species, including Achromobacter ruhlandii, Achromobacter clinical sources, were species other than A. xylosoxidans.In piechaudii, Achromobacter denitrificans, Achromobacter spanius, the current study, we sought to validate the use of nrdA Achromobacter insolitus and Achromobacter marplatensis,are sequence as a means to differentiate Achromobacter species inhabitants of soil and have been much less frequently associated and to determine the distribution of these species among with human infection [2–5]. Achromobacter-infected CF patients in the U.S. We recently developed a multi-locus sequence typing (MLST) scheme that, in addition to clearly distinguishing 2. Methods these seven Achromobacter species, identified an additional 14 novel genogroups, each of which is genetically distinct 2.1. nrdA analysis enough to represent a novel species [6]. In that study, we Amplification and DNA sequencing of nrdA was performed ⁎ Corresponding author at: Department of Pediatrics and Communicable Diseases, as described [6]. Sequences were trimmed to 765 bp and aligned University of Michigan Medical School, 8323 MSRB III, SPC 5646, 1150 W. Medical Center Drive, Ann Arbor, MI 48109-0646, USA. Tel.: +1 734 936 9767; using MegAlign (DNASTAR, Madison, WI, USA) with Clustal fax: +1 734 764 4279. W to generate a dendrogram with 1000 bootstrap replications E-mail address: [email protected] (J.J. LiPuma). using the default parameters in MegAlign. All nrdA sequences 1569-1993/$ -see front matter © 2012 European Cystic Fibrosis Society. Published by Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jcf.2012.10.002 T. Spilker et al. / Journal of Cystic Fibrosis 12 (2013) 298–301 299 can be found at the PubMLST site at http://pubmlst.org/ from the full MLST scheme as we described previously [6]. In that achromobacter [7]. For validation of distinctness of sequence analysis, only genogroup 5 could not be shown to be distinct; similarity clusters, the k parameter was calculated [8].This but, this may be a function of the next closest group (genogroup metric represents the distinctness of populations, expressed 7) containing only two strains. It is not clear whether or not k as a ratio (k) of the expected divergence between populations parameters would support distinct groups with the addition of to the expected divergence within a population. The expected more strains. Nevertheless, all three groups segregate into divergence within populations is determined largely by the distinct clusters in the three trees presented in the Supplemental diversity-purging effect of periodic selection, which is in turn Information. determined by the rate of recombination within populations Analysis of nrdA sequences demonstrated that A. xylosoxidans and the intensity of the periodic selection. The k parameter is accounted for only 143 (42%) of Achromobacter infection in our calculated as the ratio of the mean inter-group sequence cohort of 341 CF patients (Fig. 1). A. ruhlandii accounted for divergence to the mean intra-group sequence divergence. A k 23.5% of Achromobacter infection, while 17% of the strains parameter greater than 2 indicates that a group is distinct from were genogroup 14. We identified no individuals infected with its closest neighbor. A. marplatensis, A. piechaudii, or genogroups 1, 9, 10, 11 or 12. 2.2. Species distribution 4. Discussion A review of isolates referred to the Burkholderia cepacia Our previous work developing a genus-level MLST scheme for Research Laboratory and Repository (University of Michigan, Achromobacter, based on loci from seven housekeeping genes USA) identified 341 CF patients from whom Achromobacter had demonstrated several novel genogroups in this genus [6]. been recovered from respiratory specimens. Isolates had been Phylogenetic analyses provided compelling evidence that these referred from 86 CF treatment centers in the U.S. between 1998 genogroups represented novel Achromobacter species; this is and 2012. A single isolate from each patient was included in this supported by on-going comprehensive taxonomic studies, includ- study. In cases where multiple Achromobacter isolates were ing DNA–DNA hybridizations (manuscripts in preparation). In a available from the same patient, the first isolate received was similar recent study, Ridderberg et al. [9] employed multilocus used. Fifty-five of the 341 isolates had previously been analyzed sequence analysis (MLSA) based on five housekeeping genes by the recently described MLST scheme [6].Fortheremaining (with only one, rpoB, in common with our MLST scheme) to study 286 isolates, only nrdA was amplified and sequenced. a collection of 77 Achromobacter strains, including type strains of the seven recognized species, as well as environmental and clinical 3. Results strains originating in Europe, Asia, and South America. They too were able to differentiate the current Achromobacter species and As expected, among the set of 55 isolates for which both nrdA identified four novel genogroups. sequence and MLST analyses had been performed, fewer nrdA In our analyses, and as reported by Ridderberg et al. [9], alleles (n=39) were detected than were MLST sequence types sequence analysis of the bacterial 16S rRNA gene was inadequate (n=55). in differentiating novel and established Achromobacter species. A dendrogram based on the full complement of concatenat- However, in developing our MLST scheme, we noted that another ed MLST loci (2249 bp) from 47 Achromobacter strains locus, nrdA, appeared capable of clustering strains into species- (representing strains from each species, genogroup, or subgroup) level groups. This was supported by the comparison of a was compared to the dendrogram based on nrdA sequence alone dendrogram based on a 765 bp internal nrdA fragment with a (765 bp) for these same 47 strains (Supplemental Fig. 1). The tree composed of the 2249 bp concatenated MLST sequences. separation of strains into distinct groups in the nrdA tree matched Although, as expected, the topology of the trees differed, since the separation of strains into the 21 species and genogroups found nrdA sequence alone is more limited in providing phylogenetic in the concatenated MLST tree. Bootstrap values supported this inferences relative to the full MLST sequence data, both trees clustering in the nrdA sequence tree. A dendrogram with clustered strains into groups representing all 21 Achromobacter bootstrap support of clusters showing all unique nrdA sequence species and genogroups. Phylogenetic analyses, including k types identified among isolates from the 341 CF patients is parameter calculation and bootstrapping analyses, supported the provided in the Supplemental Information. utility of the nrdA locus in differentiating most groups. Increasing Similar to the results we obtained previously by using the numbers of strains analyzed may increase the k parameters for concatenated MLST sequences [6], k parameter calculations the three groups with low levels of distinctness based on this single indicated that nrdA sequence alone separated most of the 21 species locus. and genogroups into genetically distinct groups; k parameters Having established that nrdA sequence analysis provides for ranged from 2.5 to 40.9 for these groups (Supplementary data species-level separation