Lineage, Antimicrobial Resistance and Virulence of Citrobacter spp

Abstract

Citrobacter spp. are opportunistic human pathogens which can cause nosocomial infections, sporadic infections and outbreaks. In order to determine the genetic diversity, in vitro virulence properties and antimicrobial resistance profiles of Citrobacter spp., 128 Citrobacter isolates obtained from human diarrheal patients, foods and environment were assessed by multilocus sequence typing (MLST), antimicrobial susceptibility testing and adhesion and cytotoxicity testing to HEp-2 cells. The 128 Citrobacter isolates were typed into 123 sequence types (STs) of which 101 were novel STs, and these STs were divided into five lineages. Lineages I and II contained C. freundii isolates; Lineage III contained all C. braakii isolates, while Lineage IV and V contained C. youngae isolates. Lineages II and V contained more adhesive and cytotoxic isolates than Lineages I, III, and IV. Fifty-one of the 128 isolates were found to be multidrug-resistant (MDR, ≥3) and mainly distributed in Lineages I, II, and III. The prevalence of quinolone resistance varied with Lineage III (C. braakii) having the highest proportion of resistant isolates (52.6%), followed by Lineage I (C. freundii) with 23.7%. Seven qnrB variants, including two new alleles (qnrB93 and qnrB94) were found with Lineage I being the main reservoir. In summary, highly cytotoxic MDR isolates from diarrheal patients may increase the risk of severe disease.

Keywords: Citrobacter; adhesion; cytotoxicity; multidrug resistance; sequence types.

Figure 1

Phylogenetic relationships of the 128 Citrobacter isolates from this study. The phylogenetic tree of the 128 Citrobacter isolates was constructed using the concatenated sequences of the seven housekeeping genes by the neighbor-joining algorithm. Salmonella LT2 was used as an outgroup. Lineage divisions were marked. Bootstrap values of 50% or more from 1000 replicates were shown.

Figure 2

Phylogenetic relationships of the 123 C. freundii isolates from this study and two previous studies [10,16]. Lineages are marked on the node with roman numerals. Bootstrap values from 1000 replicates are shown on or near the nodes if ≥50%. The presence of ESBL and qnr genes, source, year, NDR (number of drugs resistant to), adhesion, LDH and antibiotic resistance phenotype of an isolate is shown on the right. The tree was constructed using the neighbor joining method. ST, D, F, E, H, and LDH denote sequence types, isolates from diarrheal patients, foods, environment and healthy individuals, and lactate dehydrogenase, respectively. Adhesion index: ***, >50; **, >1 and <50; *, <1; +/-, ambivalent or no adhesion; -, no adhesion.

Figure 3

Phylogenetic relationships of the 57 C. youngae isolates from this study and our previous study [10]. Lineages divisions are marked on the node with roman numerals. The number in bracket after strain name denote number of strains for ST39 which includes AH2007022, AH2007024, AH2007025, AH2008001, AH2008002). Bootstrap values (numbers on or near the nodes) from 1000 replicates are shown if ≥50%. The presence of ESBLs and qnr genes, source, year, NDR (the number of drugs resistant to), adhesion, LDH and antibiotic resistance phenotype of an isolate is shown on the right. The tree was constructed using the neighbor joining method. ST, D, F, H, and LDH denote sequence types, isolates from diarrheal patients, foods and healthy individuals, and lactate dehydrogenase, respectively. Adhesion index: ***, >50; **, >1 and <50; *, <1; +/-, ambivalent or no adhesion; -, no adhesion.

Figure 4

Phylogenetic relationships of the 53 C. braakii isolates (lineage III) from this study and our previous study [10]. The phylogenetic tree of the 53 C. braakii isolates was constructed using the concatenated sequences of the seven housekeeping genes by the neighbor-joining method. Bootstrap values of 50% or more from 1000 replicates were shown. The presence of ESBLs and qnr genes, source, year, NDR (number of drugs resistant to), adhesion, LDH and antibiotic resistance phenotype of an isolate is shown on the right. The tree was constructed using the neighbor joining algorithm. ST, D, F, H, and LDH indicate sequence types, isolates from diarrheal patients, foods and healthy individuals, and lactate dehydrogenase, respectively. Adhesion index: ***, >50; **, >1 and <50; *, <1; +/-, ambivalent or no adhesion; -, no adhesion.

Figure 5

Adhesion and cytotoxicity of Citrobacter isolates to the human epidermoid laryngocarcinoma (Hep-2, CCC0068) cell line. (A) Light micrographs of the adherence patterns exhibited by the 13 highly cytotoxic Citrobacter isolates, and control strains CF74 and CF72. Bar: 10μm. (B) Cytotoxicity of the 13 highly cytotoxic Citrobacter isolates was measured by the amount of the LDH released after 8 h exposure by Hep-2 cells. CF72 and CF74 were control strains. CF72 was a non-cytotoxic and non- adhesive negative control, CF74 was a highly adherent and cytotoxic positive control.

Figure 6

The percentage of adhesive and cytotoxic isolates in different lineages. (A) and (B) The percentage of high, intermediate, little or no adhesive or cytotoxic isolates based on the 128 Citrobacter isolates from this study. (C) and (D) The percentage of high, intermediate, little or no adhesive or cytotoxic isolates in different lineages based on the128 Citrobacter isolates from this study and 95 Citrobacter isolates from our previous studies [10,16]. The statistical significance between Lineages I and II or Lineages IV and V was determined by Mann-Whitney U test. *, p < 0.05; **, p < 0.01.

Figure 7

The number of adhesive and cytotoxic isolates in different lineages. The data were based on 128 isolates from this study and 95 Citrobacter isolates from our previous studies [10,16].