Characterization of β-lactam resistance in K. pneumoniae associated with ready-to-eat processed meat in Egypt

K. pneumoniae was known as a nosocomial infection that causes human diseases. It is considered as one of the food-borne pathogens as it causes septicemia and diarrhea in humans. This study aims to characterize K. pneumoniae strains isolated from ready to eat processed meat phenotypically and genetically. Three hundred and fifty ready to eat processed meat (Luncheon-meat) samples were collected. Forty-four (12.6%) K. pneumoniae strains were isolated and bio-typed, where the majority were identified to belong to biotype B1. K1 and K2 serotypes were detected and strains were classified as hypermucoviscous K. pneumoniae (HVKP) and classic K. pneumoniae (CKP) (26 and 18 isolates, respectively). The isolates were resistant to several classes of β–lactam antibiotics, ceftazidim and cefotaxime (95.5%), cefoxitin (93.2%), ertapenem (90.9%) and amoxicillin-clavulanic acid (86.4%). They were classified as extended spectrum β–lactamases (ESBLs), AmpC or carbapenemase-producers phenotypically. Eighteen β-lactamase genes were investigated by PCR. The most prominent genes were SHV (63.6%), TEM (52.2%), CTX-M15 (50%), AMPC (47.7%), CIT-M (45.5%) and VIM (43.2%). Co-detection of β–lactam resistance genes revealed 42 gene profiles. Twenty-four isolates had the complete efflux system (AcrAB-ToƖC). Besides, Integrons (I, II, III) were detected in 20 isolates. Molecular typing by ERIC-PCR showed high genetic diversity between isolates as 34 different patterns were identified. Overall, this study confirmed the hazards posed by the presence of multiple resistance genes in the same isolate and this should not be undervalued. Besides, the horizontal transfer of plasmid harboring resistance genes between isolates in food represents potential health risks for consumers in Egypt and so the control and inhibition plans are necessary.


Introduction
Klebsiella pneumoniae is a facultative anaerobic Enterobacteriaceae that presents as normal microbiota of skin, mouth and intestine. However, it is also responsible for pneumonia, urinary, and lower biliary tract diseases [1], intra-abdominal, blood stream infections, meningitis, a1111111111 a1111111111 a1111111111 a1111111111 a1111111111

Serotyping
The tested isolates were serologically examined for the presence of capsular antigens K1 and K2 by Quelling test [18]. The antigen-antibody reactions were observed microscopically.

String test
For the detection of mucoviscosity of the isolates, string test was performed as described previously [19]. The isolates were classified as hypermucoviscous K. pneumoniae (HVKP) isolates give positive results or classic K. pneumoniae (CKP) isolates give negative results.

Phenotypic detection of β-lactamases enzymes
Phenotypic detection of ESBLs in K. pneumoniae isolates was performed by a double-disc synergy test (DDST) according to CLSI guidelines [20]. For AmpC β-lactamases activity, the test was conducted as reported by Barwa et al. [14]. The isolates showing distortion of the cefoxitin inhibition zone were considered as AmpC producers, while isolates with no distortion of the inhibition zone were identified as AmpC non-producers. Modified Hodges test (MHT) was used to test for carbapenemases enzymes according to CLSI [20].  [11,[13][14][15] beside efflux pump genes (AcrA, AcrB and TOlC) [16]. In addition, two multiplex PCR tests were performed. The first one was used to detect metallo-β-lactamases genes (IMP, VIM, and KPC) [13]. The second one detected the integron genes I, II, and III [17]. Table 1 showed the primer sequence, annealing temperature, and the product size of the used primers. The amplified PCR products were analyzed on 1.2% agarose gel stained with ethidium bromide, compared with 100 bp plus DNA ladder (Thermofisher Scientific) and scanned in gel documentation and analysis system (Model Gel Doc 1.4, 1189; AccuLab, USA).

ERIC-PCR
Genomic DNA of each isolate was subjected to ERIC-PCR using the primers ERIC1 (5'-ATGTAAGCTCCTGGGGATTCAC-3') and ERIC2 (5'-AAGTAAGTGACTGGGGTGAGCG-3') and the PCR condition was followed as described elsewhere [1]. PCR products were analyzed by 2% agarose gel electrophoresis system. A similarity matrix was measured using Dice's coefficient and the resultant dendrogram was made via the unweighted-pair group method with arithmetic averages (UPGMA).

Statistical analysis
Graph-pad instate was used to analyze the results statistically. Fisher's exact and Chi-square tests were used and P-value � 0.05 was considered statistically significant.

Biochemical and molecular identification of K. pneumoniae
In this study, 350 ready to eat processed meat specimens (luncheon meat) were collected, forty-four K. pneumoniae isolates were isolated, purified, and biochemically identified in addition to identification using PCR by amplification of K. pneumoniae 16S-23S ITSD with an incidence of 12.6%. Three different biotypes profiles were found among isolates using biochemical activities (S2 Table). The profile B1 was the predominant one, as it was found in 32 isolates (72.7%) with a P-value < 0.0001. The other profiles B3 and B4 were found in seven (15.9%) and five (11.3%) isolates, respectively. No isolates were classified as biotype B2 or B5.
Hypermucoviscosity of isolates was measured using string test. It was found that 26 of the isolates were HVKP (P = 0.0109) while 18 isolates were CKP. The association of capsular serotypes and the hypermucoviscosity of isolates was investigated. It was found that K1 serotype was significantly associated with HVKP (P = 0.0364).

Phenotypic detection of β-lactamases enzymes
In this study, the isolates were investigated for the production of ESBL, AmpC and carbapenemase enzymes phenotypically. The results revealed that 22, 21, and 16 isolates were ESBL-producers, AmpC-producers, and carbapenemase-producers, respectively. One isolate (4.8%) was classified as an ESBL and AmpC co-producer. Nine isolates were AmpC-and carbapenemase

PLOS ONE
β-lactam resistance in K. pneumoniae and ready-to-eat processed meat co-producers. Seven isolates were ESBL and carbapenemase co-producers. Two isolates were not classified as any of the aforementioned classes. The correlation between the different β-lactamse classes and biotyping was studied. Biotype B1 was significantly present in ESBL-producers (P = 0.0455) and carbapenemase-producers (P<0.0001) while B2 was detected significantly in AmpC-producers (P = 0.0075). The serotype K1 was present in all classes, significantly (P<0.0001).
The co-detection of β-lactam resistance genes in the tested isolates was shown in Table 3. For example, TEM was significantly associated with GES, AmpC, FOX-1, ACT-1, and IMP. CIT-M was co-detected significantly with FOX-1 and VIM. Besides, IMP was significantly associated with VIM and CTX-M15.
Investigation of the gene profile of the tested isolates revealed that there were 42 different gene profiles among the 44 isolates as illustrated in Table 4. There was only one profile (Pr1) carried by two isolates while each of the other profiles was represented by one isolate. Two isolates harbored nine genes, 10 isolates carried four genes and five and six genes were carried by eight and six isolates, respectively. Resistance genes profiles were unique as each profile was represented by only one isolate.
Investigation of the distribution of integrons between isolates revealed that integrons (I, II, III) were detected in 20/44 isolates. Eleven, two, and four isolates were found to carry Int I, Int II, and Int III, respectively. Two isolates harbored both Int I and Int III and one isolate carried Int II and Int III. ESBL-producers were shown to be significantly associated with Int I (P = 0.033). In AmpC-producers, only 9/21 of isolates harbored Int I (three isolates), Int III (three isolates), each of Int II, Int I+Int III, and Int II+Int III was carried by one isolate. Six isolates in carbapenemase-producers harbored Int I (two isolates), Int III (three isolates), and Int II+ Int III (one isolate). Int I and Int III were present significantly in K1 serotyped isolates (P = 0.02 and 0.07, respectively). It was found that isolates carrying integrons were resistant to 5 or more antimicrobials (P< 0.0001), harbored 5 or more of resistance genes (P = 0.0256), and carried the complete efflux system (P = 0.0038).

ERIC
ERIC molecular typing gave high diversity among the tested isolates. As seen in Fig 1, ERIC-PCR classified the isolates into 34 different patterns. Pattern 1 included four isolates, Table 3. Co-detection rates of β-lactam resistance encoding genes amongst 44 K. pneumoniae isolates.

PLOS ONE
β-lactam resistance in K. pneumoniae and ready-to-eat processed meat Pr6

PLOS ONE
β-lactam resistance in K. pneumoniae and ready-to-eat processed meat patterns 3 and 13 each enclosed three isolates, and patterns 2 and 14 each contained two isolates while the other patterns, each one represented by one isolate. The number of bands varied between one and 12 bands, which ranged between 100 and >1500 bp. Distribution of ERIC patterns among isolates harboring β-lactam resistance encoding genes was illustrated in Table 3 where 42 different resistance gene profiles were distributed among 34 ERIC patterns that represent all tested K. pneumoniae isolates, where ERIC pattern 1 and 2 have the same gene profile (Pr1).

Discussion
K. pneumoniae is an opportunistic pathogen that is found not only in clinical specimens but also in food. Foodborne pathogens are widely distributed but research about K. pneumoniae as a food pathogen is infrequent. Recently, K. pneumoniae is the main cause of foodborne outbreaks in different countries [1]. In this study, K. pneumoniae isolates were isolated from ready to eat processed meat (Luncheon) with an incidence of 12.6% which indicates that contaminated food with K. pneumoniae is common in Egypt. Similar results were reported in china where the incidence of K. pneumoniae in 998 food samples was 9.9% [4]. Another study conducted in Spain isolated only 9 strains (5.6%) of K. pneumoniae from 160 vegetables [22]. One of the important virulence factors of K. pneumoniae is the production of a capsule that gives it a mucoid appearance on an agar plate beside that it protects it against phagocytosis and serum bactericidal effect [1]. In addition, classifying K. pneumoniae isolates as virulent ones are associated with serotypes K1 and K2 [23,24]. Two capsular types (K1 and K2) were identified by Quelling test. Two isolates (4.6%) were untypable (non K1/K2 serotype). Yu et al. had described that among 50 K. pneumoniae isolates, 26 (52%) were K1, 14 (28%) were K2 and 10 (20%) isolates were non K1/K2 serotype [20]. Other studies reported different prevalence rates for K1 and K2, where four capsular types (K1, K2, K20, and K54) were identified in 8 isolates only and the rest of the isolates were non K1/K2 serotypes [1]. Besides that, another study illustrated that K1 and K2 serotypes were found in 28.5% and 7.14% of the tested K. pneumoniae isolates, respectively [25].
Another virulence factor of K. pneumoniae, hypermucoviscosity, was measured using the string test. The results showed that 59% of isolates were HVKP and 41% of isolates were CKP. In contrast, Gharrah et al. [11] found that 33% of isolates were HVKP. Diverse capsular ingredients and an increased amount of capsular material have been described in hypervirulent K. pneumoniae isolates [26]. Similar to other studies that reported that HVKP isolates are associated with K1 or K2 serotype [23], the current results illustrated that K1 and K2 serotypes were found as both HVKP (59.1%; 23 and 3 isolates respectively) and CKP (36.4%; 14 and 2 isolates respectively) but significantly associated with HVKP (P = 0.0018). However, little work elucidating the role of the hypermucoviscous (HMV) phenotype in the pathogenicity of K. pneumoniae exists [11].
The extensive use of antimicrobials in the ecosystems resulted in the emergence of antimicrobial-resistant K. pneumoniae. In the current study, most of the isolates (93.2%) were resistant to cefotaxime. These isolates usually produce ESBL and show resistance to other β-lactam antibiotics [27,28]. In addition, the studied isolates show high resistance to ceftazidim, cefoxitin amoxicillin-clavulanic acid, and ertapenem, and moderate resistance level to aztreonam, cefepime, imipenem, meropenem, and ceftriaxone. There was a wide assortment of resistance patterns among isolates, where 15 antibiotypes were found. The predominant antibiotype was A5 which includes isolates resistant to 8 antimicrobials. 93.2% of isolates were resistant to five or more antimicrobials with MAR index �0.5. The presence of such resistant isolates in food represents a risk factor to the food consumers as they are considered as high-risk sources of antimicrobial contamination, also, they cause infections that can't be treated by these agents and result in an increase in the morbidity and mortality rates. The isolates that show resistance to ertapenem, meropenem, and imipenem (36.4%) are more likely to produce carbapenemase enzymes. These findings are contrary to what was reported by another study which found that 3.2% of isolates were resistant to amoxicillin-clavulanic acid, cefotaxime, and cefoxitin. Besides that, resistance to ceftriaxone was only 1.6% [1]. The high rates of antimicrobial resistance detected in this study can be attributed to the lack of strict policies that govern the use of antibiotics in Egypt.
In this study 22 isolates were ESBL-producers, 21 isolates were AmpC-producers and 16 isolates were carbapenemase-producers. Two isolates were not categorized as any of the previous classes. The percentage of K. pneumoniae isolates that produce ESBLs differs between countries. Their percent in Arabian countries is high (62.5% and 50% of isolates) as reported by Aljanaby and Alhasani [29] and Gharrah et al. [11], respectively. This is similar to the present results as 50% of isolates were categorized as ESBL-producers. This may have a significant effect on the treatment strategies using β-lactam antibiotics which increase the morbidity and mortality among food consumers. AmpC enzymes are β-lactamases that hydrolyze penicillins, cephalosporins, and cephamycins with low hydrolysis rates for cefepime, cefpirome, and carbapenems [30]. They aren't inactivated by available β-lactamases inhibitors in contrast to ESBLs [31]. The present results showed that 47.7% of isolates were AmpC-producers with 4.8% classified as ESBL and AmpC co-producer. Barwa et al. showed that 31.6% of isolates were AmpC-producers [14], moreover, 65 K. pneumoniae isolates were ESBL-producers and 7.7% of them were AmpC-producers as reported by Zorgani et al. [32].
Among used antibiotics, carbapenem is considered the choice for the treatment of critical infections due to multidrug-resistant K. pneumoniae. However, there is an increasing incidence of carbapenem resistance through K. pneumoniae isolates in many countries due to the extensive use of the carbapenems [33]. In the present work, 38.6% of isolates were found to be carbapenemase-producers, 56.3% of them were AmpC-coproducer while the rest were ESBL-producers too. Japoni-Nejad et al. showed that 12 isolates were carbapenemases-producers and 8 of them were AmpC-producers [34]. Besides, Yazgan et al. showed that 51% of collected K. pneumoniae were ESBL-producers, 49% of them were carbapenemase-producers [8]. These results may be attributed to the excessive use of these antibiotics. There is a significant increase in carbapenem resistance K. pneumoniae isolates in Egypt as it increased from 13.9% to 44.4% [25,35]. Several mechanisms are responsible for this type of resistance including AmpC or ESBL production with porin loss, carbapenemase production, or metallo-β-lactamase production [36].
All phenotypic methods used in this study detects isolates as positive ESBL, AmpC, or carbapenemase producer but they were unable to differentiate the different types or families of each class. Therefore, it is necessary to use other techniques such as PCR for confirmation of phenotypic results and discrimination of different genes. Examination using PCR classified 41, 35, and 31 isolates as ESBL, AmpC, and carbapenemase producers, respectively. Sutandhio et al. showed that 12 isolates harbored carbapenemase genes but only 6 isolates of them were Modified Hodges test positive [37]. Another study showed that 37.5% of K. pneumoniae isolates were ESBL or AmpC-producers phenotypically while 43% of them were ESBL only, 11% were AmpC and 36% were ESBL and AmpC genotypically [38]. Similar results were observed in other studies [14,39]. This variation between phenotypic and PCR results may be because these genes are present but are not usually expressed [14]. Therefore, not all isolates detected as ESBL, AmpC, and carbapenemases producers by PCR can be detected by phenotypic methods.
PCR is considered the gold standard method for the detection of different classes of β-lactamase enzymes [40]. The PCR amplification of β-lactam resistance genes was performed on all isolates. For ESBL genes, the most predominant were SHV, TEM, and CTX-M15 as harbored by 63.6%, 52.3%, and 50%, of isolates, respectively. Similarly, Hasani et al. illustrated that SHV gene was the predominant gene (80.9%) in the isolates followed by CTX-M and TEM (73% and 58%, respectively) [41]. For AmpC genes, AmpC, CIT-M were the most abundant genes as were carried by 47.7% and 45.5%, respectively. In contrast, Ghonaim and Moaety showed that CIT and MOX genes were present in 18.9% and 6.1%, respectively, and ACC, FOX, and ACT genes were not detected [42]. In Carbapenemases genes, VIM and KPC were the highest (43.2%) and the lowest (9.1%) genes detected respectively. In contrary to other studies that reported that OXA-48, NDM, and VIM were the predominant genes in this order [26]. ESBL and AmpC-coproducer (isolate No. 39) harbors 9 genes (Pr41). For efflux pump genes, AcrA, AcrB and TOlC were amplified in 95.5%, 63.6% and 84.1% of isolates, respectively. The coexistence of several genes of ESBL, AmpC, and carbapenemases in the same isolate revealed serious epidemiological, clinical and public health threats.
The association between β-lactam resistance encoding genes was investigated amongst the 44 K. pneumoniae isolates as illustrated in Table 3 to determine if there was a non-random association between genes that may be due to genes co-location. Some combinations of β-lactam resistance encoding genes were detected significantly. In contrast to our results, Lalzampuia et al., [43] found that for eight isolates, 7 harbored CTX-M-1 gene and/or TEM gene. The CTX-M and TEM were the most common gene combinations (33.33%) in E. coli isolates from broiler farms in the Philippines [44]. Benmahmod et al. reported that one isolate had both IMP and NDM, and one isolate co-carried IMP and VIM [45].
The investigation of the β-lactam resistance genes profile revealed that 42 different profiles were distributed among 44 K. pneumoniae isolates. The gene profile of isolates showed that there is a great genetic diversity between isolates and that the plasmids encoding for β-lactamases can be easily spread by horizontal transfer among Enterobacteriaceae including K. pneumoniae [33]. Efflux pump systems in K. pneumoniae include AcrAB and mdtK systems, that belong to (RND) and (MATE) family efflux pumps, respectively. The AcrAB-TolC pump is composed of an outer-membrane channel (TolC), a secondary transporter located in the inner membrane (AcrB), and a periplasmic component (AcrA) [25,46]. In this study, the Efflux pump system consisting of AcrAB-ToƖC was determined by PCR. Twenty-four isolates (54.5%) harbored the three genes. On the other hand, 20 isolates were missing either the AcrA or AcrB efflux pump or the ToƖC outer membrane protein. Other studies showed that 82.14% of isolates harbored AcrAB while only 5 isolates showed an incomplete AcrAB efflux system [25]. In addition, another study showed that 100% of isolates carried AcrAB while 96% carried ToƖC genes [23]. Multidrug efflux system (AcrAB-ToƖC) is responsible for the resistance of K. pneumoniae to β-lactam, tetracyclines, and quinolones [23].
Investigation of integrons distribution among isolates revealed that Integrons (I, II, III) were detected in 45.5% of isolates, of which 55%, 10%, and 20% of isolates carried Int I, Int II, and Int III, respectively. Both Int I and Int III were co-founded in 20% of isolates and 10% of isolate co-carried Int II and Int III. A study conducted by Sedighi et al. showed that Int I was present in 8% of isolates while Int II and Int III were absent from all isolates [47]. Rezai et al. reported that Int I was detected in 79.3% while Int II was harbored by 10.3% of isolates [48]. ESBL-producers were significantly associated with Int I (P = 0.033), similar to what reported by Elsherif et al., as Int 1 was associated with CTX-M gene (P = 0.039) [49]. In contrary to what reported about the association of Int III with GES gene [49], current results showed that they coexist only in one isolate out of seven Int III positive isolates. Additionally, IMP gene coexisted with Int III in 42.9% of isolates which was unlike the study conducted by Elsherif et al. who reported that Int III is of low occurrence with IMP gene [49].
In this study, the characterization of NDM-encoding isolates was examined where NDM can break down all the β-lactam antibiotics except aztreonam. This gene is located on a transmissible plasmid and is associated with other resistant genes. This may result in a broad drug resistance, which makes the treatment options limited [33]. Herein, these isolates were significantly classified as K1 (P = 0.0011) and B1 (P<0.0001). Half of them were classified as HVKP. Only four isolates harbored integrons, three of them carry Int I and one isolate carry IntII+-IntIII. Nine out of the 12 isolates were confirmed as carbapenemase-producers by phenotypic and PCR methods, while three isolates were identified only by PCR. This may be attributed to one of the Modified Hodges test drawbacks in that it gives false-negative results for NDM producers [50]. NDM gene was present as sole carbapenemase gene in four (33%) isolates, significantly associated with VIM gene in six isolates (P<0.0001), with IMP gene in three isolates (25%), with OXA-48 in four isolates (33%) and with KPC in two isolates (16.7%)(P = 0.0249), in addition, it was significantly associated with ESBL genes (GES and OXA-1-like) with P = 0.0011 and <0.0001, respectively. Khalil et al. also reported the association between NDM and other carbapenemase genes as KPC, IMP, VIM, and OXA-48 where they were detected in NDM-1 producing K. pneumoniae with the frequency of 80%, 30%, 50%, and 30% respectively [33]. Eight isolates harbored the three efflux genes (66.7%) with P = 0.0214. Although NDM is a broad spectrum carbapenemase that can inactivate β-lactam except for aztreonam [25], four isolates in this study showed resistance to aztreonam and harbored NDM indicating that a new antimicrobial resistance pattern in Egyptian isolates may exist.
Recently, ERIC-PCR is one of the most potent tools used for the analysis of genetic relatedness in K. pneumoniae [33]. ERIC molecular typing results gave high diversity among the tested isolates as it classified the isolates into 34 different patterns. The association between phenotypic and molecular typing gave no clear relation between isolates except in the case of isolates No. 47 and 41among ESBL-producers and isolates No. 22 and 19 among carbapenemase-producers. The relation between ERIC patterns and β-lactam resistance encoding genes profiles was studied where 42 different resistance gene profiles were distributed among 34 ERIC patterns that represent all tested K. pneumoniae isolates. Similar results were reported by Kazemian et al. [51] Our results indicated the significance and convenience of ERIC-PCR and β-lactam resistance encoding genes profiles in differentiating isolates based on genetic relatedness. The diversity in our isolates represented a problem in the treatment of K. pneumoniae infections. Similar genetic diversity represented by ERIC was described in previous studies [9,33].
In conclusion, the present study indicated that food (ready to eat processed meat; Luncheon) is a good reservoir of resistant K. pneumoniae isolates. This represents a public health problem and good control of the emergence and the transmission of these isolates is needed. This can be done by developing more prevention strategies on making and selling this food in supermarkets.
Supporting information S1