Virulence of Klebsiella pneumoniae Isolates Harboring bla KPC-2 Carbapenemase Gene in a Caenorhabditis elegans Model

Klebsiella pneumoniae carbapenemase (KPC) is a carbapenemase increasingly reported worldwide in Enterobacteriaceae. The aim of this study was to analyze the virulence of several KPC-2-producing K. pneumoniae isolates. The studied strains were (i) five KPC-2 clinical strains from different geographical origins, belonging to different ST-types and possessing plasmids of different incompatibility groups; (ii) seven transformants obtained after electroporation of either these natural KPC plasmids or a recombinant plasmid harboring only the bla KPC-2 gene into reference strains K. pneumoniae ATCC10031/CIP53153; and (iii) five clinical strains cured of plasmids. The virulence of K. pneumoniae isolates was evaluated in the Caenorhabditis elegans model. The clinical KPC producers and transformants were significantly less virulent (LT50: 5.5 days) than K. pneumoniae reference strain (LT50: 4.3 days) (p<0.01). However, the worldwide spread KPC-2 positive K. pneumoniae ST258 strains and reference strains containing plasmids extracted from K. pneumoniae ST258 strains had a higher virulence than KPC-2 strains belonging to other ST types (LT50: 5 days vs. 6 days, p<0.01). The increased virulence observed in cured strains confirmed this trend. The bla KPC-2 gene itself was not associated to increased virulence.


Introduction
Klebsiella pneumoniae carbapenemases (KPC) are broad-spectrum b-lactamases, which confer resistance to all b-lactam molecules, including carbapenems. Those plasmid-encoded enzymes are identified mostly in K. pneumoniae [1][2][3]. KPC producers have been increasingly reported worldwide [1]. Eleven KPC variants (KPC-2 to KPC-12) are known differing by a few amino acids changes, KPC-2 being the most frequent KPC-type enzyme. KPC-associated enterobacterial infections do not seem to be specific to sites, organs, or tissues: most are either systemic infections, occurring in patients with multiple invasive devices or urinary tract infections without an indwelling catheter, particularly in immunocompromised patients [4,5]. No specific virulence factors have been identified as being associated to KPC-producing strains [1]. Risk factors associated with the acquisition of KPCproducing bacteria included prolonged hospitalisation, intensive care unit hospitalisation, invasive devices, immunosuppression, and multiple antibiotic agents before initial culture as known for many other multidrug resistance Gram-negative bacilli [4,6].
The present study was aimed to evaluate the virulence of several bla KPC-2 producing K. pneumoniae strains using Caenorhabditis elegans model of infection.

Bacterial Strains
The studied K. pneumoniae strains are listed in Table 1. Five clinical KPC-2-producing K. pneumoniae strains were studied (origin of isolates and MLST type are indicated in brackets) ( Table 1) [7][8][9][10][11]. Virulence of these strains for C. elegans was compared with K. pneumoniae strain ATCC10031/CIP 53153. E. coli strain OP50, an avirulent strain corresponding to food for lab worms, was used as the negative control in the C. elegans model. Bacterial strains were cultured onto Trypticase soya (TS) agar plates at 37uC under a 5% CO 2 atmosphere. Bacterial isolates were stored at -80uC.
The construction of the recombinant plasmid pRYC-1 that contained a bla KPC-2 gene into cloning vector pBKCMV has been previously described [11]. This recombinant plasmid was extracted from the recipient strain, purified and electroporated into competent K. pneumoniae ATCC10031/CIP 53153. Transformants were selected on Luria Broth agar plates containing 0.5 mg/L of imipenem. The cloning vector pBKCMV which did not contain a bla KPC-2 gene was also electroporated into competent K. pneumoniae ATCC10031/CIP 53153 using kanamycin as selection agent.

Plasmid Curing
The curing of plasmids of each clinical K. pneumoniae strain was performed by successive bacterial subcultures at 44uC as described previously [12]. The loss of the plasmids was visualized by agarose gel electrophoresis and confirmed by negative PCR results for the selected plasmid genes including bla KPC-2 gene (Table 1).

Resistance Traits
Susceptibility to b-lactams in particularly to carbapenems, fluoroquinolones, tigecycline, cotrimoxazole, aminoglycosides, and colistin was determined by either the E-test (bioMérieux, Marcy l'Etoile, France) or by the agar disk diffusion method. MIC determination was performed and interpreted as recommended by the Clinical and Laboratory Standards Institute (CLSI), as updated in 2012. The bla KPC , bla CTX-M , bla SHV , bla TEM , bla OXA-1/9 -like genes were searched by PCR and characterized as described previously [1,11].
The presence of a capsule was determined by staining with India ink. On a slide, a drop of bacterial suspension was mixed with the stain. By light microscopy, the capsule appears as a clear area between the colored background and the stained bacterial body.

Nematode Killing Assay
The C. elegans model has been developed to study host-pathogen interactions. In this test, the studied bacteria is presented as food to   the nematodes instead of E. coli strain OP50, an avirulent strain, that is their usual food in the lab. Ingestion of the bacteria studied by the worms results in an infection and ultimately the killing death of the worms [14]. The time required for bacteria to kill the worms compared with life duration observed when the worms are fed with E. coli strain OP50, is an indirect marker of virulence potential of the studied bacteria. The C. elegans infection assay was carried out as described by Lavigne et al. [14] using the Fer-15 mutant line, which has a temperature sensitive fertility defect. To synchronize the growth of worms, eggs were collected using the hypochlorite method. Overnight cultures in nematode growth medium (NGM) of K. pneumoniae strains were harvested by centrifugation, washed once and suspended in phosphate buffered saline solution (PBS) at pH 7.0 at a concentration of 10 5 CFU/ml. NGM agar plates were inoculated with 10 ml of each strain and incubated at 37uC for 8 to 10 h. Plates were allowed to cool to room temperature and seeded with L4 stage worms (20-30 per plate). Plates were then incubated at 25uC and scored each day for live worms under a stereomicroscope (Leica MS5) [14]. At least three replicates repeated five times were performed for each selected clone. Lethal time 50% (LT50) corresponded to time (hours) required to kill 50% of the nematode population. The results are representative of five independent assays for each group of strains.

In vitro Strain Growth
Bacteria were grown in LB broth and NGM broth. An aliquot (1 mL) of an overnight culture of each strain was transferred in 99 mL and cultured at 37uC under shaking conditions. Bacterial growth was monitored by change of the optical density at 620 nm (OD 620 ) and determined over a period of 24 h. Growth curves were performed in triplicate for each strain and the OD 620 measurements were averaged for each time point in order to plot growth curves.

Bacterial Ingestion in Worms
In vivo experiments consisted of measuring the number of bacteria within the C. elegans digestive tract 72 h after ingestion as described by Garsin et al. [15]. Five C. elegans were picked at 72 h, and washed twice in 4 ml drops of M9 medium. The nematodes were placed in a 1.5 ml Eppendorf tube containing 20 ml of M9 medium with 1% Triton X-100 and were mechanically disrupted by using a pestle. The volume was adjusted to 50 ml with M9 medium containing 1% Triton X-100 which was diluted and plated on TS agar containing 50 mg/L of ampicillin. At least three replicates were performed for each bacterial challenge. Antibiotic susceptibility of each strain recovered from the digestive tract of the nematodes was determined by the disk agar diffusion method and compared to that of the corresponding strain given as food to the nematodes.

Bacterial Lawn Avoidance Assay
Small lawns of the different studied strains were cultured on NGM plates overnight at 25uC. Approximatively 30 young adult worms grown on OP50 were put in the centre of each bacteria lawn. The number of animals on each lawn was counted after 4, 8 and 16 h as previously described [16]. The results were explained in percent occupancy corresponding to the number of worms in the bacterial lawn/the total number of worms. The experiments were performed in triplicate.

Statistical Analysis
In order to perform pairwise comparison between two different strains, we used a log rank test. The Kruskall-Wallis test was used to compare the in vitro and in vivo bacterial growth of the different strains. A p value #0.05 given by SPSS 6.1.1 (SAS Institute Inc, Cary, NC, USA) was considered as reflecting statistical significance.

Antibiotic Susceptibility
As with most clinical KPC producers [1], the strains were susceptible to colistin and tigecycline. K. pneumoniae KN2303 and K. pneumoniae KN633 were susceptible to all tested aminoglycosides and fluoroquinolones, K. pneumoniae YC was susceptible only to gentamicin, K. pneumoniae A28006 was susceptible to aminoglycosides, and K. pneumoniae 475 was only susceptible to amikacin, levofloxacin and ciprofloxacin.
The isolates contained several plasmids differing in structure and size. In each case, a single plasmid harboring the bla KPC-2 gene was electroporated into the same K. pneumoniae reference strain ATCC10031/CIP 53153. The antibiotic susceptibilities of the transformants are shown in Table 2. PCR and plasmid extraction analysis confirmed the presence of a single plasmid harboring a bla KPC-2 gene in those transformants. Transformants were resistant to at least ceftazidime and carbapenems, which was related to the expression of the carbapenemase KPC-2 ( Table 2). In most of the cases, co-resistance to non-ß-lactam antibiotics was not observed for those transformants. The only exception was for transformant p475, which showed resistance to aminoglycosides (gentamicin and tobramycin) and cotrimoxazole. Conversely, strains cured of plasmids had lost resistance to extended-spectrum cephalosporins and carbapenems (data not shown).

Virulence
The distribution of virulence factors is shown in Table 1. The different adhesins-encoding genes were present in all the isolates. The siderophore gene entB was also present in 100% of the strains. However, only the reference strain ATCC10031/CIP53153 harbored aerobactin genes (irp1, irp2, ybtS). None of our isolates carried rmpA, magA, cnf1, and hlyA genes.
The results of nematode killing assays are presented in Table 3.
The clinical KPC-producing strains were significantly less virulent (LT50 varied between 5.0 to 6.0 days 60.5) than reference strain ATCC10031/CIP 53153 (LT50: 4.3 days 60.2) (p,0.001), but significantly more virulent than the avirulent OP50 (LT50: 8.2 days 60.3) (p,0.001). Nematode killing results allow us to constitute two groups of clinical KPC strains. A first group included K. pneumoniae YC, 475 and A28006 (ST-258 type and related ST type) with a mean survival time of 5 days (60.5) and a second group of K. pneumoniae KN2303 and KN633 (non-ST-258 type), with a mean survival time of 6 days (60.5). The difference in virulence between both groups was significant (p,0.01). Then, the five plasmids (pBC633, pBC2303, pNYC, pA28006 and p475) were electroporated into competent K. pneumoniae ATCC10031/ CIP 53153. Comparison between the virulence of clinical strains and the corresponding transformants showed no significant difference (p = 0.4, 0.10, 0.96, 0.81 and 0.13 for YC, A28006, 2303, 633 and 475, respectively). The three transformants (pNYC, p475 and pA28006) strains were significantly more virulent than that of the two transformants (pBC633 and pBC2303) strains (LT50 varied between 5.5 to 5.6 vs 5.9 to 6.3 days 60.5, respectively) (p,0.001) showing an increased virulence strains harboring KPC plasmids from strains belonging to ST-type 258.
The five transformant strains remained significantly less virulent than the recipient ATCC10031/CIP 53153 strain (LT50 varied between 5.5 and 6.3 vs 4.3 days 60.5, respectively) (p,0.001) suggesting that the introduction of the plasmid-mediated bla KPC-2 gene decreased the virulence of the reference strain (Table 3). To support the direct effect of the plasmid in the bacterial virulence, the virulence of plasmid cured strains was determined. The cured strains were all significantly more virulent than the corresponding parental strains (LT50 varied between 4.3 to 5.0 vs 5.0 to 6.0 days 60.5, respectively) (p,0.001).
To analyze the role of the bla KPC-2 gene itself, virulence of a same strain (ATCC10031/CIP 53153) harboring a recombinant plasmid containing the bla KPC-2 gene (pRYC-1) was analysed and compared to that of the same plasmid without the bla KPC-2 gene. Strain harboring pRYC-1 was significantly less virulent than the reference strain and the three most virulent clinical strains (YC, 475 and A28006) (LT50 varied between 6.6 vs 5.5 and 5.6 days 60.5, respectively) (p,0.001). In addition, introduction of the plasmid pBK-CMV (cloning plasmid) without the bla KPC-2 gene did not modify the virulence of the strain (LT50 varied between 4.6 vs 4.3 days 60.3, respectively) (p = 0.19) ( Table 3).

In vitro Bacterial Growth
The growth curves determined in LB medium did not show significant difference between the studied strains and the corresponding transformants. The same results were obtained by culturing the strains in the NGM medium which is used for the nematode assays (data not shown). Thus, the presence or introduction of a KPC-encoding plasmid into reference strain had no impact on the in vitro bacterial growth. This also confirms that the in vivo results were not biased by growth in NGM medium.

Bacterial Ingestion
The differences in killing were not due to differences in the survival and proliferation of the strains in the nematode intestine since the bacterial colonization with the different strains in the digestive tract of nematodes measured 72 h after ingestion was 10 6 CFU without significant difference (data not shown). The antibiotic susceptibility of the strains obtained after passages in the digestive tract was strictly identical to that of the strains given as food to worms (data not shown).
Moreover, the occupancy of the worms on bacterial lawns was similar for all the strains around 100% without significant difference (Table 4).

Discussion
Recent data suggest a relationship between K. pneumoniae virulence and drug resistance [17][18][19][20]. Expression of efflux pumps in K. pneumoniae as a mechanism of antibiotic resistance may be associated with expression of novel virulence factors required to resist to innate immune defense mechanisms in nematode and murine models and may be also associated to resistance to the host antimicrobial peptides [17,18]. The impact of carbapenemase expression on virulence is weak [21]. A recent study suggests that known virulence factors such as K1, K2, and K5 capsular polysaccharides, rmpA and the aerobactin gene were absent in KPC-producing isolates and these strains present low virulence in a murine lethality model [21]. In this present study, we have investigated the influence played by different plasmids harboring a single carbapenemase gene, the bla KPC-2 gene on the bacterial ability to colonize and infect the C. elegans nematode. Five clinical isolates from US, Colombia, Israel and Brazil which belonged to different ST clones were studied. In nematode killing assay, the KPC-2 clinical isolates were significantly less virulent (LT50: 5 to 6 days 60.5) than reference strain ATCC10031/CIP 53153 (LT50: 4.3 days 60.2) (p,0.001). The analysis of the distribution of virulence factors in each group of strains could explain this observation with the absence of some siderophores genes in KPC isolates. However, we found the presence of the genes encoding adhesins in all the strains that demonstrates a virulence potential of the resistant strains (as also suggested by the results of the strains with cured plasmids). Some other virulence factors are now well established in K. pneumoniae strains. K antigen, a capsular polysaccharide, is considered to be an important virulence factor of K. pneumoniae. According to these capsular polysaccharides, K. pneumoniae can be classified into 77 serological K antigen types. K1 isolated in South East Asia was the predominant serotype causing pyogenic liver abscess. K2 serotype is among the most common capsule types isolated from patients with UTI, pneumonia, or bacteremia. It can be assumed that K2 is the predominant serotype of human clinical isolates worldwide [22,23]. In KPC clinical strains, these serotypes are absent limiting the effect of this important factor of virulence [21].
Among the KPC-2 clinical strains, two groups of bacterial virulence can be constituted: the first group including clone ST-258 and its single-or double-locus variants, ST-11 and ST-277 with an higher virulence trait than the second group including clones ST-377 and ST-378 (p,0.01). Recent findings indicate the international dissemination of KPC-producing strains of ST-258 type [24]. We showed here that an increase of virulence of KPC-2 producing strains belonging to ST-258 (compared to non-ST-258 strains) may be due to the plasmid type from those KPC-2 producers of the ST-258 type. However if we compared the strains 475 and KN2303 each belonging to one of the different ST groups, they display LT50 pattern while they share the same plasmid type (IncN). This implies that other factors than just plasmids are probably also involved in virulence of KPC-2 producers in this model.
In the aim to complete our investigation, we studied the role of the bla KPC-2 gene in KPC-2 positive plasmids contained in ST-258 K. pneumoniae. Actually its presence decreased virulence of the recombinant strains. These results may be explained either by the natural role of the bla KPC-2 gene itself or a modulation of expression of chromosome-encoded genes. Some bias due to the fact that the bla KPC-2 gene was carried on a high-copy vector could not be excluded. Therefore, we believe that other factors expressed by plasmids of the KPC-2 plasmids of ST-258 type strain, such as colonization fitness, survival capability in the nosocomial environment, adherence, biofilm production may explain their increased virulence or/and dissemination. Indeed the nucleotide sequence of plasmids carrying the bla KPC-2 gene is known in several cases [25,26]. Different virulence factors have been detected notably in strains of ST-258 type, which harboured a type 4-secretion system gene cluster [25]. The impact of this secretion system inducing a short, rigid pilus releasing virulence factors must be studied to evaluate its real role in K. pneumoniae virulence.
In conclusion, this study emphasizes that the key point with regard to the balance between virulence and resistance in the multidrug resistant bacteria studied in this work is at least the type of KPC-2 plasmids.