Extensive Expression of the Virulome Related to Antibiotic Genotyping in Nosocomial Strains of Klebsiella pneumoniae

The emergence of hyper-virulent and multidrug-resistant (MDR) strains of Klebsiella pneumoniae isolated from patients with hospital- and community-acquired infections is a serious health problem that increases mortality. The molecular analysis of virulome expression related to antimicrobial-resistant genotype and infection type in K. pneumoniae strains isolated from patients with hospital- and community-acquired infections has been poorly studied. In this study, we analyzed the overall expression of the virulence genotype associated with the antimicrobial resistance genotype and pulse field gel electrophoresis (PFGE) type (PFtype) in K. pneumoniae. We studied 25 strains of K. pneumoniae isolated from patients who developed bacteremia and pneumonia during their hospital stay and 125 strains from outpatients who acquired community-acquired infections. Susceptibility to 12 antimicrobials was determined by Kirby–Bauer. The identification of K. pneumoniae and antibiotic-resistance genes was performed using polymerase chain reaction (PCR). To promote the expression of the virulence genes of K. pneumoniae, an in vitro infection model was used in human epithelial cell lines A549 and A431. Bacterial RNA was extracted with the QIAcube robotic workstation, and reverse transcription to cDNA was performed with the Reverse Transcription QuantiTect kit (Qiagen). The determination of the expression of the virulence genes was performed by real-time PCR. In addition, 57.3% (n = 86) of the strains isolated from patients with hospital- and community-acquired infections were multidrug-resistant (MDR), mainly to beta-lactam antibiotics (CB, AM, CFX, and CF), aminoglycosides (GE), quinolones (CPF and NOF), nitrofurantoin (NF), and sulfamethoxazole/trimethoprim (SXT). The most frequently expressed genes among strains isolated from hospital- and community-acquired infections were adhesion-type, ycfm (80%), mrkD (51.3%), and fimH (30.7%); iron uptake, irp2 (84%), fyuA (68.7%), entB (64.7%), and irp1 (56.7%); and protectins, rpmA (26%), which were related to antibiotic-resistance genes, blaTEM (96%), blaSHV (64%), blaCITM (52.6%), blaCTXM-1 (44.7%), tetA (74%), sul1 (57.3%), aac(3)-IV (40.7%), and aadA1 (36%). The results showed the existence of different patterns of expression of virulome related to the genotype of resistance to antimicrobials and to the PFtypes in the strains of K. pneumoniae that cause hospital- and community-acquired infections. These findings are important and may contribute to improving medical treatment strategies against infections caused by K. pneumoniae.

The pathogenic capacity of K. pneumoniae is associated with the expression of a large number of virulence markers, which include adhesins (fimbriae Type 1 and Type 3), ironacquisition systems, protectins, and toxins [15].The genes that encode the fimbriae (fimH-1, kpn, ycfM, and mrkD) are considered the main virulence markers of K. pneumoniae that mediate adhesion to host cells.Therefore, their expression favors bacterial colonization [16], while the expression of siderophore genes, such as enterobactin (entB), yersiniabactin (irp1, irp2, fyuA, and ybtS), salmochelin (iroN) and aerobactin (iutA), allowing for iron uptake and facilitating survival and bacterial multiplication during infections [17], while the expression of the cnf-1 and hlyA toxin genes promote the degradation of the host tissues, facilitating bacterial dissemination [18].
The increase in mortality rates in patients with hospital-and community-acquired infections caused by hyper-virulent and multidrug-resistant strains of K. pneumoniae is a critical health situation [19], so it is currently important to analyze the properties molecular virulence and antimicrobial resistance of the new variants of K. pneumoniae with the purpose of improving medical treatment alternatives.
Molecular characterization of hyper-virulent hospital strains of MDR K. pneumoniae in Mexico [20,21], and in other parts of the world, have been studied [22][23][24].However, reports on the expression of virulence markers, and their correlation with multidrug-resistance and clinical origin, are scarce.Therefore, in this study, we aimed to establish an in vitro model of infection using human cell lines to globally and comprehensively determine the virulome expression related to the antimicrobial resistance genotype and pulse field gel electrophoresis (PFGE) type in a group of K. pneumoniae strains from different infections.

Multidrug-Resistance Phenotype
Utilizing the Kirby-Bauer disk diffusion method, our analysis revealed notable levels of antibiotic resistance among the tested strains, with the highest percentages of resistance observed for carbenicillin, ampicillin, and cefotaxime, as well as for nitrofurantoin (Table 1).The 57.3% (n = 86) of the K. pneumoniae strains isolated from patients with hospital-and community-acquired infections were multidrug-resistant in the range of 4-12 antimicrobials, within which more than 40% (n = 60) of the strains presented the phenotype of multi-resistance to nine antibiotics from different groups; beta-lactams (CB, AM, CFX, and CF), aminoglycosides (NET), quinolones (CPF and NOF), nitrofurans (NF), and sulfonamide/trimethoprim (SXT).

Origin of the Strains and In Vitro Infection of Human Epithelial Cell Lines to Determine Viruloma Expression
The majority of hospital-acquired K. pneumoniae strains were isolated from patients with bacteremia (21/150; Table 2) and community-acquired infections from patients with UTIs (61/150) and respiratory infection (53/150).Using the real-time PCR method, we found that the most frequently expressed genes among the strains after the in vitro infection model of the human epithelial cell lines A549 and A431 were the genes encoding adhesins (ycfm, mrkD, and fimH), the genes encoding iron-acquisition systems (irp2, fyuA, entB, and irp-1) and protectins (rpmA; Table 2).The frequency of expression of most virulence genes in the strains according to clinical origin was very similar, except for adhesin (mrkD) and enterobactin (entB), whose percentages were higher in pneumonia and bacteremia (hospital-acquired infections), respectively, while yersiniabactin (ybtS) and protectin (rpmA) percentages were higher in others (community-acquired infections).The genes expressed with the least frequency after the in vitro model of infection of the cell lines with K. pneumoniae strains were hlyA (n = 6) and magA (n = 1), both from strains isolated from patients with infections acquired in the community.The gene that codes for the CNF-1 toxin was not expressed in any of the strains studied.

Detection of Antibiotic Resistance Genes by Conventional PCR
The most frequently detected antibiotic-resistance genes by the conventional PCR method among K. pneumoniae strains were beta-lactams (bla TEM , bla SHV , bla CITM , and bla CTXM-1 ), tetracycline (tetA), sulfonamide (sul1), gentamicin (aac(3)-IV), streptomycin (aadA1), and trimethoprim (dfrA1).No statistically significant differences were found between the frequency of most antibiotic-resistance genes in strains according to clinical origin, except for bla CTX-1 and aac(3)-IV, where frequencies were higher in strains isolated from bacteremia and pneumonia of patients with hospital acquired infections than in others (Table 3), respectively.The antibiotic resistance genes bla CTXM-2 (n = 4) and qnr (n = 2) in strains isolated from patients with hospital-acquired infections (UTI and respiratory infection) had the lowest frequency.

Virulence Genes Expression According to Antibiotic Resistance Genes
Overall, in the K. pneumoniae strains isolated from patients with hospital-acquired infections, our real-time PCR assays showed an elevated expression of adhesion genes (mrkD and ycfM) and iron-acquisition systems (entB, irp1, irp2, and fyuA) associated with the presence of antibiotic-resistance genes to beta-lactam antibiotics (bla SHV , bla CITM , bla TEM , and bla CTXM-1 ), sulfonamides (sul1), and tetracycline (tetA) (Table 4).However, statistically, the individual frequency of expression of each virulence gene related to the detection of each of the antibiotic-resistance genes in the strains was different (Table 4).
Similarly, among the K. pneumoniae strains isolated from patients with communityacquired infections (Table 5), we detected a high expression by real-time PCR of adhesion genes (mrkD and ycfM), and iron acquisition systems (entB, irp1, irp2, and fyuA) related to detection by conventional PCR of resistance genes to beta-lactam antibiotics (bla SHV and bla TEM ) and tetracycline (tetA).However, statistically, the individual frequency of expression by real-time PCR of each virulence gene related to the detection by conventional PCR of each of the antibiotic-resistance genes in the strains was different (Table 5).

Unsupervised Hierarchical Clustering
Unsupervised hierarchical clustering analysis showed three main groups based on similarities between K. pneumoniae strains (Figure 2).Group 1 (strain range 40-108) had the highest number of expressed virulence genes and was characterized by the simultaneous expression of one-to-seven virulence genes.Group 2 (strains between 70-147) had an intermediate number of expressed virulence genes (four to nine genes), and Group 3 (strains between 138-48) had the lowest number and was defined by the expression of 3-10 virulence genes.

Discussion
The emergence of hypervirulent and MDR hospital-and community-acquired strains of K. pneumoniae is a serious concern that increases mortality, mainly in immune-compromised patients [14,25].Numerous virulence factors contribute to the pathogenicity of K. pneumoniae, thus increasing the severity of hospital and community infections [26].Our working group previously determined the frequency of virulence genes by conventional PCR in these K. pneumoniae strains (n = 150), as well as its relationship with the hypermucoviscosity phenotype, capsular serotypes, biofilm formation, and the resistance phenotype to antibiotics [27].In this work, we continue our research by characterizing in these K. pneumoniae strains the expression of the virulome by real-time PCR related to the antimicrobial resistance genotype, to the pulse field gel electrophoresis (PFGE) type, and their clinical origin.
In this study, the overall percentages for the expression of different virulence genes in K. pneumoniae strains isolated from patients with hospital-and community-acquired infections following the in vitro infection of human epithelial cell lines were very similar, except for the frequency of expression of mrkD, entB, ybtS, and rmpA genes, where there were statistically significant differences.We found a higher expression of the adhesion Based on the expression of the virulome (VE) by real-time PCR, two different clades were identified: (A) Highly expressed among the strains (ymfM, irp-2, fyuA, entB, irp-1, and mrkD), and (B) Lowly expressed between strains (fimH, iutA, rpmA, kpn, ybtS, iron, hlyA, magA, and cnf ).The antibiotic-resistance genotype (ARG) was grouped into two main clades according to their frequency of detection by conventional PCR.Clade C had a large number of genes detected (bla TEM , tet(A), sul-1, bla SHV , bla CTXM-1 , bla CITM , aadA1, and aac(3)-IV), and Clade D was composed by a lower number of gene (dfrA-1, cmlA, cat1, tet(B), bla CTXM-9, bla CTXM-2, and qnr).
Overall, the unsupervised hierarchical analysis showed in the three groups (1, 2, and 3) strains with the same pattern of virolome expression.For example, in Group 2, it is observed in the cladogram that Strains 101, 98, 90, 25, 9, and 21 presented the same expression profile of the virulence genes (ycfM, irp-1, fyuA, entB, irp-1, and mrkD) associated with the frequency of the genotype and phenotype of resistance to antibiotics bla TEM -TET(A)-AM-CB, and with the clinical origin.Interestingly, this shared pattern of expression was independent of the strain origin because Strains 101, 98, 90, and 9 were isolated from bacteremia of patients with hospital-acquired infections, as well as Strains 25 (infected ulcer) and 21 (respiratory infection) from patients with community-acquired infections.The Strains 113-50-79; 45-7-43 from Group 1, and 102-134, 42-104, and 2-12 from Group 3, also had the same virulome expression pattern.

Discussion
The emergence of hypervirulent and MDR hospital-and community-acquired strains of K. pneumoniae is a serious concern that increases mortality, mainly in immune-compromised patients [14,25].Numerous virulence factors contribute to the pathogenicity of K. pneumoniae, thus increasing the severity of hospital and community infections [26].Our working group previously determined the frequency of virulence genes by conventional PCR in these K. pneumoniae strains (n = 150), as well as its relationship with the hypermucoviscosity phenotype, capsular serotypes, biofilm formation, and the resistance phenotype to antibiotics [27].In this work, we continue our research by characterizing in these K. pneumoniae strains the expression of the virulome by real-time PCR related to the antimicrobial resistance genotype, to the pulse field gel electrophoresis (PFGE) type, and their clinical origin.
In this study, the overall percentages for the expression of different virulence genes in K. pneumoniae strains isolated from patients with hospital-and community-acquired infections following the in vitro infection of human epithelial cell lines were very similar, except for the frequency of expression of mrkD, entB, ybtS, and rmpA genes, where there were statistically significant differences.We found a higher expression of the adhesion marker mrkD in strains from hospital-acquired infections (pneumonia) than in strains from community-acquired infections.The adhesion gene ycfM was expressed in all strains isolated from bacteremia (21/21), pneumonia (4/4), and stool culture and tumor biopsy (3/3; other).The co-expression of the adhesion genes fimH, mrkD, and ycfM detected in some strains of K. pneumoniae, following the infection of A549 and A431 cell lines, suggests that this expression combination could increase bacterial colonization and persistence in infected tissues.Type 1 fimbriae (fimH) was previously found to be frequently expressed during the infection of urinary tissue and participate in the invasion of bladder cells [16], while Type 3 fimbriae (mrkD) was found to promote biofilm formation and, consequently, the evasion of the host immune response [28].
Iron is an essential element for electron transport, DNA synthesis, and peroxide reduction, favoring bacterial survival and multiplication during infections [29].Among the expression of iron uptake genes, entB expression was higher in bacteremia and pneumonia (hospital-acquired infections), and in respiratory infection (community-acquired infection) as well.High expression percentages of irp1 (bacteremia, respiratory infection, and others), irp2, and fyuA (bacteremia, UTI, respiratory infection, and others) were also found.The simultaneous high expression of the iron uptake genes entB, irp1, irp2, ybtS, and fyuA by some of the strains after infection of A549 and A431 cell lines highlights the virulence of the strains to cause acute infections because entB (enterobactin) and irp-1 (yersiniabactin) expression increases biofilm formation in Klebsiella pneumoniae, causing liver abscess [30]; ybtS (yersiniabactin) promotes respiratory tract infection [31], while irp and fyuA (yersiniabactin) reduce the bactericidal capacity of innate immune cells [32].
The expression of the mucoid phenotype A (rmpA) gene was detected mainly in the strains from respiratory infection and UTI.This proves the ability of the strains to cause more acute infections because rmpA has been previously associated with hvKpn strains, causing liver abscesses and bacteremia [33], while hlyA, which was more frequently expressed in strains originating from UTI (community-acquired infection), favors bladder inflammation during UTIs [34].
More than half of the K. pneumoniae strains (n = 86) isolated from patients with hospitaland community-acquired infections were multidrug-resistant (MDR), mainly to beta-lactam antibiotics (CB, AM, CFX and CF), aminoglycosides (GE), quinolones (CPF and NOF), nitrofurantoin (NF), and sulfamethoxazole/trimethoprim (SXT), which coincides with the multidrug-resistance described in strains of K. pneumoniae isolated from patients with hospital-acquired infections in Mexico [20].
The frequency of the detection of antibiotic-resistance genes in K. pneumoniae strains isolated from patients with hospital-and community-acquired infections was very similar, except for bla CTXM-1 (beta-lactam) and aac(3)-IV (gentamicin), where the frequency was higher in strains isolated from bacteremia and pneumonia, respectively, both from hospitalacquired infections.High percentages for beta-lactam (bla SHV and bla TEM ), tetracycline (tet(A)), and sulfonamide (sul1)-resistance genes were found in strains isolated from patients with hospital-acquired infections (bacteremia and pneumonia) and community-acquired infections (UTI, respiratory infection, infected ulcer, and others).High percentages for the streptomycin-(aadA1) and gentamicin-(aac(3)-IV) resistance gene were also detected in strains from patients with hospital-acquired infections (pneumonia) and communityacquired infections (UTI, respiratory infection, and infected ulcer).The overall number of bla TEM (144/150) and bla CXTX-9 (10/150) coincide with those described in K. pneumoniae strains isolated from Mexican patients in a health facility [35], while the numbers of bla SHV (96/150), bla CTXM-1 (67/150), and bla TEM (144/150) are higher than those described in K. pneumoniae strains isolated from hospital-acquired infections in other parts of the world [36,37].The wide distribution of the resistance genotype to other non-beta-lactam antibiotics, such as tetracycline (tet(A)), sulfonamide (sul1), gentamicin (aac(3)-IV), and streptomycin (aadA1) detected in K. pneumoniae strains from patients with hospital-and community-acquired infections coincides with the high MDR phenotype found in Mexico in hospital strains of K. pneumoniae from bacteremia, UTIs, respiratory infection, and other infections [35,38].
In this study, the correlation of the expression percentages of adhesion genes (mrkD and ycfM) and iron-acquisition systems (entB, irp-1, irp2, and fyuA) with the genotype of resistance to beta-lactam antibiotics (bla SHV , bla TEM , and bla CTXM-1 ), sulfonamide, and tetracycline was higher in K. pneumoniae strains isolated from patients with hospital-acquired infections compared with strains from community-acquired infections.These results are substantial and suggest that the expression of the virulence genotype related to the antimicrobial-resistance genotype in strains from bacteremia and pneumonia (hospitalacquired infections) may cause more acute infections, increasing patient mortality.Mortality in patients with hypervirulent K. pneumoniae pneumonia has been found to be 23.1% [39], 54.3% in bacteremia, 48.9% in intensive care, 43.1% in organ transplantation, and 13.5% in urinary tract infection [40].The prevalence of the cnf-1 gene (n = 22) and the magA gene (n = 25), which had been previously identified by our research group within the genome of these K. pneumoniae strains [27], does not correspond with the expression results in this study using an in vitro infection model in A549 and A431 cell lines with the same strains.Interestingly, cnf-1 was not expressed in any of the strains, whereas magA was expressed in only one strain.This observed inconsistency in detection and expression is probably attributable to the cell lines used in our study since prior reports have demonstrated differential levels of expression according to the infected cell type.One example is that cnf-1 tends to have a higher frequency of expression during urinary tract infections [41], whereas magA expression has been linked to hypermucoviscosity in strains associated with liver abscesses [42].Furthermore, the temporal dynamics of the infectious process have a profound effect on gene expression.Some genes are known to be expressed early in the infection, while others have later expression.For instance, the early-expressed type 1 fimbrial genes are regulated by an invertible element located in the promoter region [43].This element permits gene expression to be repressed or activated based on the prevailing conditions of the infection process.
PFGE is a frequently used method for clonal typing of isolates in outbreaks and allows for source identification [44].In this study, 13 different PFtypes distributed in Subtypes B1 and B2 of Clade B were identified, some of them integrated by two, three, four, and up to eight strains.In Clade A, 19 PFtypes distributed in Subtypes A1 and A2 were identified.These comprehensive findings showed different virulome expression patterns associated with the antibiotic-resistance genotype and clinical origin in K. pneumoniae strains from the same PFtype, where the largest PFtype (subtype A2) was found to comprise nine strains (12,61,73,76,94,101,112,124, and 133) isolated from hospital-acquired (bacteremia [n = 3]) and community-acquired infections (respiratory infection [n = 3], and infected ulcers [n = 2] and UTI [n = 1]), all with different virulence gene expression profiles associated with the antibiotic resistance genotype.
The results showed that K. pneumoniae strains from the same PFtype, isolated from different patients with hospital or community infections, are possibly found in the hospital environment.Therefore, some patients with community-acquired strains could have acquired these strains during their hospital stay.
The unsupervised hierarchical clustering analysis, similar to the PFGE method, revealed a diverse spectrum of virulome expression patterns associated with the genotypephenotype characteristics of antibiotic resistance and the origin of K. pneumoniae strains isolated from patients with either community-or hospital-acquired infections.High throughput technologies, such as whole genome sequencing, could provide a more comprehensive understanding of the global composition of the virulence and antimicrobial-resistance genes.These analyses will be performed in future works.To the best of our knowledge, this is the first study performed in Mexico where the wide distribution of different virulome expression patterns associated with the antibiotic resistance genotype in K. pneumoniae strains isolated from patients with different hospital-and community-acquired infections is shown, suggesting the high pathogenic potential of these strains during infections.

Origin of the Strains
A total of 150 strains of K. pneumoniae were analyzed, which were collected from September 2019 to March 2020 in the Microbiology laboratory of the Hospital General Regional No. 72 (Instituto Mexicano del Seguro Social, State of Mexico, Mexico) located in the municipality of Tlalnepantla de Baz, Edo. from Mexico, Mexico.The strains were isolated from samples of hospitalized patients with ongoing infections, such as bacteremia (n = 21) and pneumonia (n = 4), which they acquired during their hospital stay after admission for managing complications arising from other comorbidities such as diabetes, high blood pressure, chronic obstructive pulmonary disease, and obesity.Additionally, strains were collected from the non-hospitalized external community, encompassing UTIs (urinary tract infection; n = 61), respiratory (n = 53), ulcers (n = 8) and other infections (stool culture (n = 2), and tumor biopsy (n = 1)).All patients provided informed consent for inclusion in this study.The study was approved by the institution's Ethics Committee (identification code: R-2022-1406-033).

Bacterial DNA Extraction
The DNA of the K. pneumoniae strains was extracted by the boiling method described previously [45].Each of the strains was seeded on eosin methylene blue agar (EMB; DIBICO, Edo. de México, Mexico) at 37 • C for 12 h under constant agitation.Four isolated 2-mm (in diameter) colonies were collected, suspended in Eppendorf tubes with 200 µL of sterile water, incubated at 100 • C for 10 min, and centrifuged at 10,000× g for 5 min.DNA obtained in the supernatant was stored at −20 • C. In order to use 100 ng/µL DNA for the PCR reaction, DNA concentration and purity were measured using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA).4.3.Identification of K. pneumoniae through Polymerase Chain Reaction (PCR) K. pneumoniae strains were identified using PCR amplification of 16S-23S rDNA internal transcribed spacer [46].The final volume per reaction mixture for each uniplex PCR assay was 20 µL; 1 µL of forward primer and 1 µL of reverse primer (10 pmol, Integrated DNA Technologies, San Diego, CA, USA), 12 µL of Taq DNA Polymerase 2× Master Mix RED (AMPLIQON, Copenhagen, Denmark), 3 µL of nuclease-free water, and 3 µL of DNA template (100 ng).The PCR amplification conditions were performed using a T100™ Thermal Cycler (Bio-Rad, Feldkirchen, Germany) instrument as follows; initial denaturation at 94 • C for 5 min, 30 cycles of denaturation at 94 • C for 30 s, annealing at 55 • C for 30 s, and extension at 72 • C for 50 s; with a final extension of 72 • C for 5 min.The K. pneumoniae ATCC 700721 strain was used as a positive control.The amplified DNA fragments (130 pb) were visualized using 2% agarose gels stained with Midori Green Direct (Nippon Genetics, Düren, Germany).

Bacterial Dilution of Strains for Infection of Human Epithelial Cell Lines
One colony (2 mm in diameter) of every K. pneumoniae strain was seeded separately in brain heart infusion broth (MCD Lab, Edo. de México, Mexico) and incubated for 12 h at 37 • C under constant shaking.Following the instructions in the RNA protect Bacteria Reagent Manual (Qiagen, Hilden, Germany), a 1:4 dilution of each K. pneumoniae culture was performed using phosphate-buffered saline solution to obtain an optical density at 600 nm = 1.0 using a Beckman DU-7400 spectrophotometer (Laguna Hills, CA, USA), which corresponded to a concentration of 1 × 10 9 cells/mL.Subsequently, dilutions were performed in order to obtain a concentration of 2 × 10 6 cells/mL.

Preparation of Human Epithelial Cell Lines and In Vitro Infection Model
The in vitro infection model using human epithelial cell lines to determine virulence gene expression was performed as previously described [50,51].With the aim of emulating a human infection process, we prepared cultures of human lung cancer A549 (ATCC CCL-185) and human epidermoid A431 (CRL-1555, Manassas, VA, USA) cell lines.The cell lines were plated in DMEM medium (Dulbecco's Modified Eagle Medium, Corning NY, USA) and supplemented with 10% FBS (Fetal Bovine Serum, Corning NY, USA) in a 24-well cell culture plate, until a confluency of 1.8 × 10 5 cells/well was reached for the in vitro infection experiment.All cultures were incubated at 37 • C under 5% of CO 2 atmosphere.
To stimulate the bacterial infection of the human cell lines, each dilution of K. pneumoniae culture isolated from patients with bacteremia, pneumonia, and respiratory, ulcer, and other infections were inoculated (50 µL; 2 × 10 6 cells/mL) onto the surface of a monolayer of the cultured 1.8 × 10 5 cells of the epithelial cell line A549 (ATCC CCL-185) derived from lung cancer, while dilutions (50 µL; 2 × 10 6 cells/mL) of K. pneumoniae from patients with UTIs were inoculated on the monolayer of the human epidermoid cell line A431 (CRL-1555, Manassas, VA, USA).The 24-well plates were incubated with 1 mL of F12K plus 10% fetal bovine serum at 37 • C for 48 h under a 5% CO 2 atmosphere with saturated humidity.The maintenance medium (F12K plus 10% fetal bovine serum) was changed every 24 h.The K. pneumoniae strain, ATCC 700721, was used as a control.

K. pneumoniae RNA Extraction and Reverse-Transcription to cDNA
To establish the expression of virulence genes, RNA was extracted from K. pneumoniae strains after in vitro infection of human epithelial cell lines, for which K. pneumoniae strains were harvested from the surface of A549 and A431 cell line cultures and suspended in 1000 µL of RNA Protect Bacteria reagent (Qiagen).Samples were centrifuged at 8000× g for 10 min to obtain bacterial cell sediment.RNA was extracted from the bacterial sediment using a QIAcube automated extraction equipment (Qiagen, Hilden, Germany) with the commercial RNeasy Mini Kit (Qiagen), which involved bacterial lysis with TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 8) containing 1 mg/mL lysozyme.A NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA, EE.UU) was used to determine the concentration and purity of total RNA.The Reverse Transcription QuantiTect kit (Qiagen) was used to perform first-strand cDNA synthesis.

Determination of K. pneumoniae Virulome Expression Using Real-Time PCR
To determine virulome expression in K. pneumoniae strains by real-time PCR, the Rotor-Gene Q 5plex HRM device (Qiagen, Hilden, Germany) was used.The primers used to assess the expression of the virulence genes of K. pneumoniae virulence genes encoding for adhesins, fimH-1 (type 1 fimbriae), mrkD (type 3 fimbriae), kpn (fimH-type adhesin), and ycfM (outer membrane lipoprotein); iron-acquisition systems, entB (enterobactin biosynthesis), irp1, irp2, ybtS (yersiniabactin biosynthesis), fyuA (yersiniabactin receptor), iutA (aerobactin receptor), and iroN (catecholate siderophore receptor); protectins, magA (mucoviscosity-associated gene A) and rmpA (mucoid regulator phenotype A), and toxins; hlyA (hemolysin) and cnf-1 (necrotizing cytotoxic factor 1) were as previously described [15].The final volume-perreaction mixture for each real-time PCR assay was 20 µL: 12.5 µL of the Rotor-Gene SYBR Green PCR kit master mix (Qiagen), 1 µL of each forward and reverse primer (1 µM), 8.5 µL of RNAase-free water, and 2 µL of the cDNA (100 ng).Amplification was performed at 95 • C for 5 min, followed by 40 cycles of 95 • C for 5 s, and an extension at 60 • C for 10 s.For each quantitative real-time PCR (qPCR) assay, we performed a standard curve prepared from three dilutions of cDNA (100, 200, and 300 ng/µL) from a strain harboring the gene of interest (positive control).From these three cDNA dilutions, the Rotor-Gene Q 5plex HRM System software version 2.3.1.49(Qiagen) calculated the cutoff for the definition of the threshold cycle (TC) of each strain.In this way, with the CT values of each strain obtained with respect to the CT of the standard curve of the control strain, the arithmetic mean was obtained, and the expression percentage of each gene was determined.Each real-time PCR assay included a melting curve, a housekeeping gene (rpoB), and a non-templated control (NTC).The K. pneumoniae strains from clinical isolates Kp1, Kp6, and Kp12, harboring the 15 virulence genes studied (fimH-1, mrkD, kpn, ycfM, entB, irp1, irp2, ybtS, fyuA, iutA, iroN, magA rmpA, hlyA, and cnf-1) were used as positive controls for the preparation of the standard curve.

Pulse Field Gel Electrophoresis (PFGE)
The preparation of agarose blocks with K. pneumoniae genomic DNA was performed using the method previously described by the standardized PFGE protocol PulseNet, USA [52].To separate XbaI fragments, electrophoresis was performed on 1% of agarose gels and 0.5X TBE buffer at 14 • C with an increased pulse time from 2.2 to 54.2 s for 20 h and 6.0 V/cm using the CHEF-MAPPER (Bio-Rad, Hercules, CA, USA) device.XbaI fragment sizes were estimated using XbaI fragments of Salmonella enterica serotype Braenderup global standard H9812.The Gel Logic 112 image photo documentation system (Kodak, Rochester, NY, USA) was used to digitize the images.The BioNumerics v.7.1 software package (Applied Maths, Sint Martens-Latem, Belgium) was used to analyze the fingerprint profile on the PFGE gel.After background subtraction and gel normalization, the fingerprints in the profiles were typed based on band similarity and dissimilarity, using Dice's similarity coefficient and the unweighted pairwise clustering method with arithmetic mean according to average linkage clustering methods.PFGE patterns were interpreted according to the criteria previously described for the typing of bacterial strains [53].4.11.Unsupervised Hierarchical Clustering K. pneumoniae strains were systematically grouped according to virulome expression, antibiotic-resistance genotype, antibiotic-resistance phenotype, diagnosis, and clinical origin from patients with hospital-and community-acquired infections using unsupervised hierarchical clustering based on Euclidean distances for categorical variables.A categorical data matrix that included virulence gene expression, antibiotic-resistance genotype, antibiotic-resistance phenotype (resistance and sensible), and the clinical origin of patients with hospital-acquired infections and community-acquired infections, as well as diagnosis, was created in R (v.3.6.2).The distance of each strain was calculated based on the overall similarity coefficient, which estimates the maximum possible absolute discrepancy between each combined pair of strains.Strains were visualized in a distribution diagram with a dendrogram constructed using ComplexHeatmap (v3.6.2,R core).

Statistical Analysis
The χ 2 test was applied using the SPSS statistical software (version 20.0; SPSS Inc., Chicago, IL, USA) (p < 0.05) to establish differences between the frequency of expression of virulence genes related to the antibiotic-resistance genotype and clinical origin in K. pneumoniae strains isolated from patients with hospital-and community-acquired infections.

Conclusions
Our results highlighted the existence of different PFtypes of K. pneumoniae causing hospital-and community-acquired infections, with wide virulome expression profiles related to the antimicrobial-resistance genotype and infection type.Therefore, it is important to implement new monitoring and control programs on the emergence of new hypervirulent and MDR K. pneumoniae strains to improve medical treatment and reduce mortality rates.

Figure 1 .
Figure 1.Pulse-field gel electrophoresis (PFGE) profile dendrogram and expression patterns of virulence genes and antibiotic-resistance genotyping in K. pneumoniae strains from hospital-and community-acquired infections.The dendrogram was produced through the Dice similarity coefficient and unweighted pairwise clustering method with arithmetic mean clustering methods using PFGE imaging of XbaI-digested genomic DNA.The scale bar shows the correlation coefficient (%).Capital letters (A,B) represent the two main clades produced.The letters B1, B2, A1, and A2 represent subgroups of strains.The strain number, type of infection, origin, and virulence gene expression pattern associated with the antibiotic resistance genotype are shown on the right.

Figure 2 .
Figure 2. Hierarchical grouping of K. pneumoniae strains according to virulome expression related to antibiotic-resistance genotype-phenotype, diagnosis, and clinical origins.Positivity and negativity for a particular genotype is represented by a red and grey rectangle, respectively.Upper panel: Virulome expression (VE).Middle panel: Antibiotic-resistance genotype (ARG).Lower panel: Antibiotic-resistance phenotype (ARP).Left axis: Expression of virulence factors (Clades A and B) and detected antibiotic-resistance genes (Clades D and C).Right axis: Absolute frequency of expression by gene and antibiotic resistance.Top: Cladogram total of VE, ARG, and ARP of the strains.Bottom: Diagnosis and origin of the strains.ARP section: AM = ampicillin, CB = carbenicillin, NF = nitrofurantoin, CF = cephalothin, SXT = trimethoprim-sulfamethoxazole, CFX = cefotaxime, GE = gentamicin, CPF = ciprofloxacin, NOF = norfloxacin, NET = netilmicin, AK = amikacin and CL = chloramphenicol.

Figure 2 .
Figure 2. Hierarchical grouping of K. pneumoniae strains according to virulome expression related to antibiotic-resistance genotype-phenotype, diagnosis, and clinical origins.Positivity and negativity for a particular genotype is represented by a red and grey rectangle, respectively.Upper panel: Virulome expression (VE).Middle panel: Antibiotic-resistance genotype (ARG).Lower panel: Antibioticresistance phenotype (ARP).Left axis: Expression of virulence factors (Clades A and B) and detected antibiotic-resistance genes (Clades D and C).Right axis: Absolute frequency of expression by gene and antibiotic resistance.Top: Cladogram total of VE, ARG, and ARP of the strains.Bottom: Diagnosis and origin of the strains.ARP section: AM = ampicillin, CB = carbenicillin, NF = nitrofurantoin, CF = cephalothin, SXT = trimethoprim-sulfamethoxazole, CFX = cefotaxime, GE = gentamicin, CPF = ciprofloxacin, NOF = norfloxacin, NET = netilmicin, AK = amikacin and CL = chloramphenicol.
(1) Identical: Strains are designated as genetically indistinguishable if their restriction patterns have the same number of bands.(2) Closely related: The PFGE patterns of the strains present two or three bands of difference.(3) Possibly related: The PFGE pattern of the strains differs between four and six bands.(4) Unrelated: The PFGE pattern differs in seven or more bands.

Table 1 .
Antibiotic resistance phenotype in strains of Klebsiella pneumoniae isolated from patients with hospital-and community-acquired infections.

Table 2 .
Frequency of expression of virulence genes in strains of Klebsiella pneumoniae isolated from patients with hospital-and community-acquired infections.

Table 3 .
Frequencies of antibiotic-resistance genes in strains of Klebsiella pneumoniae isolated from patients with hospital-and community-acquired infections.

Table 4 .
Distribution of virulence genes expression according to antibiotic resistance genes in strains of Klebsiella pneumoniae isolated from patients with hospitalacquired infections.

Table 5 .
Distribution of virulence genes expression according to antibiotic resistance genes in strains of Klebsiella pneumoniae isolated from patients with communityacquired infections.