Introduction

Acinetobacter baumannii is a clinically important pathogen responsible for a wide range of hospital-acquired infections [1]. The ampC gene of A. baumannii was cloned and sequenced for the first time in 2000 [2]. The gene, also called bla_ADC for Acinetobacter-Derived Cephalosporinase, is intrinsic in A. baumannii and all other members of the Acinetobacter calcoaceticus-Acinetobacter baumannii (Acb) complex [3, 4]. It is located in the chromosome between folE, encoding a GTP cyclohydrolase I enzyme, and an open reading frame encoding a hypothetical protein, as seen in the A. baumannii reference strain ATCC 17978-mff (GenBank accession number CP012004, locus tag ACX60_05710). Overexpression of ampC, due to the acquisition of a strong promoter located on an insertion sequence (IS) element, is the main mechanism of resistance to third-generation cephalosporins in A. baumannii [5]. With few exceptions, variation in the amino acid sequence of AmpC in A. baumannii usually does not affect the resistance spectrum [6, 7].

Some A. baumannii isolates were reported to carry a second copy of the ampC gene, located elsewhere in the chromosome [8, 9]. The additional copy was part of a DNA segment most likely derived from the chromosome of another A. baumannii strain. The segment was mobilized as part of Tn6168, a composite transposon made of two directly oriented copies of ISAba1 [8]. The A. baumannii ampC gene, together with an upstream ISOur1, was also detected in the genome of Oligella urethralis, leading to a cephalosporin resistance phenotype [10]. Interestingly, A. baumannii strain ACICU, from global clone 2 (GC2), was found to carry a 9 kb chromosomal segment, containing ISAba125-ampC, which was derived from a GC1 isolate [11]. This finding indicated the occurrence of a replacement in the chromosome of ACICU, most likely mediated by a homologous recombination event [11]. Similarly, distinctive ISAba1-associated ampC alleles were detected in the genome of GC1 isolates, once again highlighting the frequent occurrence of horizontal transfer of chromosomal DNA segments in A. baumannii [9, 12].

To track these imports, a clear numbering system of the ampC alleles is needed. Analysis of the ampC locus could also be a convenient method for exploring the molecular epidemiology of A. baumannii, taking into consideration that particular ampC alleles have been linked to certain clones of A. baumannii [9, 13]. This report aims to announce the establishment of a database for the ampC locus in A. baumannii.

New database for the ampC locus in A. baumannii

The database is hosted and maintained at the pubmlst platform for A. baumannii (http://pubmlst.org/abaumannii/) sited at the University of Oxford [14]. The platform provides an open access to all the data and allows submissions of novel sequences. However, novel sequence must simultaneously be submitted and assigned accession numbers by the International Nucleotide Sequence Database Collaboration (INSDC) (http://www.insdc.org/). Sequences must be complete and meet the validation criteria of INSDC. ampC sequences with novel nucleotide identities (≥ 1 nucleotide substitution) will be numbered successively.

So far, we have identified, curated and numbered a total of 48 distinctive alleles of the ampC locus in a collection of 188 A. baumannii isolates by means of the online available whole genome sequence records (Table 1). The ampC alleles 1, 3, 4, 5, 6, 7, 8, 13, 14, 17, and 18 were carried by isolates that belong to clonal complex 1 (CC1), corresponding to GC1, according to the Pasteur scheme for multilocus sequence typing (https://pubmlst.org/abaumannii/). Isolates from CC2, corresponding to GC2, had the ampC alleles 2, 3, 19, 20, 21, 22, 23, 24, 26, 27, 28, and 46. Nonetheless, ampC allele 2 was also present in one isolate from ST215 (27, 2, 7, 2, 2, 1, 2), which was not closely related to CC2. Similarly, allele 19 was present in isolates of ST500 (3, 3, 2, 2, 28, 1, 5) or ST522 (3, 3, 89, 2, 28, 1, 5), which were also not related to CC2. Although it was present in few isolates from CC1 and CC2, allele 3 was mainly characteristic for ST3 (3, 3, 2, 2, 3, 1, 3) or ST32 (1, 1, 2, 2, 3, 4, 4). The ampC locus alleles 10 and 16 were characteristic for CC10 and ST16, respectively. Likewise, all isolates from CC25 had the ampC locus allele 25. Allele 39 was present in all the ST78 (25, 3, 6, 2, 28, 1, 29) isolates, but also in one isolate from ST241 (40, 3, 15, 2, 40, 4, 4).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Numeration of the ampC gene alleles in Acinetobacter baumannii.

https://doi.org/10.1371/journal.pone.0176695.t001

These linkages demonstrate that sequence analysis of the ampC variants is probably a practical method to search for clinically significant clones of A. baumannii, as previously described for the intrinsic bla_OXA-51-like gene [15, 16] However, the frequent occurrence of inter-strain exchanges of chromosomal segments should be taken into consideration. Therefore, analysis of ampC to study the epidemiology of A. baumannii should be complemented by characterizing other loci or preferably be taken within the context of whole-genome sequence analysis.

Updated list of the AmpC protein variants

In parallel, we revised and updated a previous collection of the AmpC variants (Table 2) [13]. As previously recommended, the AmpC variants were numbered according to the chronology of getting published and/or submitted to the INSDC databases. Numbers were preceded by a hyphen. When it was possible, numbers assigned by previous studies were retained. Accordingly, AmpC-1 was used to label the first AmpC protein variant reported in 2000 [2, 13]. The designation AmpC-72 (GenBank accession: AIL90389) was omitted since it showed 100% amino acid similarity to AmpC-70 (GenBank accession: KQG48886). Two variants with different amino acid sequences were designated as AmpC-57 (GenBank accessions: ADO51072 and AEZ36052). Subjectively, AmpC-57 was given to the variant detected in two A. baumannii isolates from East Africa [17]. New variants were defined, based on ≥ 1 amino acid substitution, and numbered under supervision of the INSDC curators. It is very important to re-emphasize that the AmpC variant numbers (Table 2) are not matching and not exchangeable with the ampC allele numbers (Table 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Numeration of the AmpC protein variants encoded by Acinetobacter baumannii.

https://doi.org/10.1371/journal.pone.0176695.t002

Concluding remarks

In our opinion, having two databases, one for the gene alleles and one for the protein variants, will create a lot of confusion. With the rapid accumulation of bacterial whole genome sequences, we argue that genes and alleles should reasonably be defined and numbered based on their nucleotide identities. For molecular epidemiological studies, the novel database for ampC in A. baumannii will provide unambiguous details beyond traditional list of AmpC variants that are limited to alleles with amino acid substitutions. To conclude, we emphasize on using the basic definition of the word “allele” for bacterial genes, by which novel alleles should be defined regardless if they are associated with amino acid changes or not.

Acknowledgments

We thank Dr. Daniel Haft at the National Center for Biotechnology Information (NCBI) for his advice and recommendations with regards to the list of AmpC protein variants.