Introduction

Klebsiella pneumoniae is a Gram-negative bacillus, usually encapsulated, nonmotile bacterium and belongs to the Enterobacteriaceae family. It is an opportunistic pathogen that can cause several infections and is among the most common nosocomial pathogens worldwide (Wyres et al. 2020). K. pneumoniae strains can contain a wide range of virulence and antimicrobial resistance factors (Ashurst and Dawson 2020). Several virulence factors further increase the severity of infections by K. pneumoniae, such as siderophores that are systems of chelating molecules that can competitively eliminate iron from host proteins and other sources; capsules, including their overproduction characterized by the hypermucoviscous (HM) phenotype; among others (Paczosa and Mecsas 2016).

The presence of clinically relevant antimicrobial resistance genes (ARGs), including extended-spectrum β-lactamases (ESBLs) and carbapenemases-encoding genes, can limit therapeutic options, making it difficult to treat infections in affected patients (Piperaki et al. 2017). Multidrug-resistant (MDR) and hypervirulent K. pneumoniae (hvKp) clones have become major global health problems (Wyres et al. 2020).

This study aimed to determine the pathogenic potential, antimicrobial resistance profile, and the clonal relationships among K. pneumoniae strains belonging to CG258 from patients admitted to hospitals in northeastern Brazil.

Materials and methods

Bacterial strains

This study included 13 K. pneumoniae strains isolated from different patients admitted to a tertiary hospital in Teresina, a Brazilian municipality located in the northeast region, being the capital of the state of Piauí, with an estimated population of 868,075 inhabitants. The strains were randomly isolated between August and September 2013 from several sources (catheter tip, hemoculture, purulent secretion, tracheal secretion, and urine) (Table 1). The strains were identified by the VITEK 2 system (bioMérieux, France) and confirmed by the 16S rRNA sequencing, after amplification and sequencing with the primers fD1 (5'-AGAGTTTGATCCTGGCTCAG-3') and rP2 (5'-ACGGCTACCTTGTTACGACTT-3'), according to Weisburg et al. (1991). The strains were maintained at − 80 °C in 15% glycerol for subsequent experiments.

Table 1 General data related to 13 K. pneumoniae strains used in this study
Full size table

Hypermucoviscosity test

The HM phenotype was investigated using the string test according to Wiskur et al. (2008). The strains were inoculated on Mueller–Hinton agar (Oxoid Ltd., United Kingdom) and incubated for 18 h at 37 °C. After bacterial growth, using a bacteriological loop, an isolated colony was touched and raised vertically. The HM phenotype was considered positive when there was the formation of a viscous filament ≥ 5 mm.

Antimicrobial susceptibility testing

Antimicrobial susceptibility testing was realized by the disk diffusion method on Mueller–Hinton agar (Oxoid Ltd., United Kingdom) following the recommendations of the Clinical Laboratory Standards Institute (CLSI 2020) and the Brazilian Committee on Antimicrobial Susceptibility Testing (BrCAST 2020). For this approach, 40 different antimicrobial disks (Oxoid Ltd., United Kingdom) recommended for Enterobacteriales were used. The susceptibility for colistin (COL) was determined by the broth microdilution method (CLSI 2020). Each strain was considered susceptible or non-susceptible (either intermediate or resistant) to each antimicrobial tested. The antimicrobials tested are shown in Fig. 1.

Fig. 1
figure 1

Frequency of non-susceptibility found among the 13 strains of K. pneumoniae studied. Each color represents a class of antimicrobials. Antimicrobials tested: amoxicillin-clavulanate (AMC), amikacin (AMI), ampicillin-sulbactam (APS), aztreonam (ATM), ceftazidime (CAZ), cefaclor (CFC), cefixime (CFM), cefoxitin (CFO), cefazolin (CFZ), ciprofloxacin (CIP), chloramphenicol (CLO), colistin (COL), cefepime (CPM), ceftriaxone (CRO), cefuroxime (CRX), ceftaroline (CTL), cefotetan (CTT), cefotaxime (CTX), doripenem (DOR), doxycycline (DOX), ertapenem (ERT), streptomycin (EST), fosfomicin (FOS), gentamicin (GEN), imipenem (IPM), levofloxacin (LEV), lomefloxacin (LMX), minocycline (MIN), meropenem (MPM), nalidixic acid (NAL), netilmicin (NET), nitrofurantoin (NIT), norfloxacin (NOR), ofloxacin (OFX), piperacillin-tazobactam (PIT), sulphonamide (SUL), trimethoprim-sulfamethoxazole (SUT), ticarcillin-clavulanate (TAC), tetracycline (TET), tobramycin (TOB), and trimethoprim (TRI)

Full size image

The strains Escherichia coli ATCC® 25922 and Pseudomonas aeruginosa ATCC® 27853 were used as controls in these experiments. Based on the susceptibility profile, the K. pneumoniae strains were classified as MDR, extensively drug-resistant (XDR), or pandrug-resistant (PDR) according to Magiorakos et al. (2012).

Detection of virulence genes and ARGs by polymerase chain reaction (PCR)

The strains were subjected to conventional PCR assays to detect K1 (magA, K1 capsule-specific wzy gene) and K2 capsular serotypes and nine gene encoding virulence factors: fimH and mrkD (adhesins), ycfM (lipopolysaccharide), entB, iutA, ybtS, and kfu (iron acquisition systems), allS (allantoin metabolism), and rmpA (capsular serotype and HM regulator phenotype) (Compain et al. 2014; Fang et al. 2004; Yu et al. 2007). Moreover, 17 ARGs were investigated, including 16 β-lactamase-encoding genes (blaGES, blaIMP, blaKPC, blaNDM, blaOXA-1-like, blaOXA-48-like, blaVIM, blaSPM, blaGIM, blaCTX-M-Gp1, blaCTX-M-Gp2, blaCTX-M-Gp8, blaCTX-M-Gp9, blaCMY-2, blaVEB, and blaBEL) and the COL resistance gene mcr-1 (Clímaco et al. 2013; Dallenne et al. 2010; Ellington et al. 2007; Liu et al. 2016; Peirano et al. 2011; Pitout et al. 2005; Poirel et al. 2010).

One amplicon from each gene found was randomly selected to confirm its identity by sequencing using an automated sequencer (ABI 3500xL Genetic Analyzer; Applied Biosystems, USA). The obtained sequences were compared with those available in GenBank using the BLAST algorithm (http://blast.ncbi.nlm.nih.gov/Blast.cgi). The ycfM gene, although confirmed by sequencing, was not deposited due to its small size amplicon (< 200 bp). In addition, all wzi genes were directly analyzed and deposited, when necessary, in the K. pneumoniae MLST database (https://bigsdb.web.pasteur.fr/klebsiella/klebsiella.html).

Determination of capsular type

The capsular type (K-type) of the strains was determined by amplification and sequencing of the wzi gene as described by Brisse et al. (2013). The wzi gene was amplified by PCR and the amplicons were sequenced using the automated sequencer (ABI 3500xL Genetic Analyzer; Applied Biosystems, USA). Subsequently, the obtained sequences were submitted to the K. pneumoniae MLST database (https://bigsdb.web.pasteur.fr/klebsiella/klebsiella.html) to determine the alleles corresponding to each K-type and K-locus of the strains.

Multilocus sequence typing (MLST)

The strains were evaluated by the MLST technique using protocol 2 of the K. pneumoniae MLST database (http://www.pasteur.fr/recherche/genopole/PF8/mlst/Kpneumoniae.html). To show the clonal relationships among sequence types (STs) and clonal groups (CGs), goeBURST 1.2.1 version (http://www.phyloviz.net/goeburst) was used.

Results and discussion

Antimicrobial resistance profiles

The majority of strains were non-susceptible to nitrofurantoin (NIT), ciprofloxacin (CIP), and trimethoprim-sulfamethoxazole (SUT) (n = 13, 92.8%), the other members of the class of folate pathway antagonists also demonstrated high percentages, such as sulphonamide (SUL) and trimethoprim (TRI) (both 85.7%). Similar results were obtained for the tetracyclines class, such as doxycycline (DOX), tetracycline (TET) (both 85.7%), and minocycline (MIN) (78.6%). All strains were susceptible to cefotetan (CTT). Almost all were susceptible to COL since the minimal inhibitory concentration (MIC) found was < 1 mg/L, except for the KpPi151 strain that showed resistance to COL (MIC 32 mg/L). However, the mcr-1 gene was not found in this strain, indicating that resistance to COL is due to other genes or mechanisms not investigated in this study, as described by El-Sayed et al. (2020). Figure 1 shows the percentage of non-susceptibility of each antimicrobial.

A significant and worrying level of multidrug resistance strains was found, being ten strains classified as MDR and three strains (KpPi150, KpPi156, and KpPi159) as XDR (Table 1). None of the strains were classified as PDR. Among these strains, seven (KpPi145, KpPi146, KpPi147, KpPi149, KpPi151, KpPi155, and KpPi156) did not present any of the investigated ARGs, suggesting that other genes or mechanisms not researched in this study may be involved, such as efflux pumps, gene target modifiers, and changes in antimicrobial targets by mutations (Blair et al. 2015). The ideal to fully characterize the ARGs among the strains would be to characterize their resistomes by whole genome sequencing, which was not the aim of the study. Many K. pneumoniae strains with multiple antimicrobial resistance have been reported around Brazil. MDR and XDR K. pneumoniae strains have been found in the north (Ferreira et al. 2019), northeast (Aires et al. 2017), southeast (Braun et al. 2018), and south (Nava et al. 2019). In a study by Gonçalves et al. (2017), a total of 26 strains of K. pneumoniae, including 14 MDR, 7 XDR, and 3 PDR strains were found in a university hospital in southern Brazil, corroborating with our results.

Presence of ARGs and the relationship with antimicrobial resistance profiles

Three different ARGs were detected among the strains, including blaOXA-1-like, blaCTX-M-Gp1, and blaKPC (GenBank accession numbers MT330307, MT330309, and MT330311). The blaOXA-1-like was detected in four strains—KpPi148, KpPi152, KpPi153, and KpPi159 (Table 1). According to Sugumar et al. (2014), the production of OXA-1 and other β-lactamases results in resistance to ampicillin, piperacillin, ticarcillin, and cephalosporins; in fact, this profile was observed in our results. Four strains (KpPi148, KpPi152, KpPi153, and KpPi159) demonstrated resistance to APS, PIT, TAC, CAZ, CFC, CFM, CFZ, CPM, CRO, CRX, CTL, CTX, and CFO, and the only exceptions were CFO for which two strains (KpPi148 and KpPi152) were susceptible and for CTT which all strains were susceptible. All tested antibiotics are shown with their abbreviations in Fig. 1.

The KpPi152 strain also has blaCTX-M-Gp1, being the unique strain to present this gene. Chagas et al. (2011) studied the diversity of genotypes in 38 CTX-M-producing K. pneumoniae from different Brazilian hospitals located in three southeastern cities and suggested that the intensive use of broad-spectrum cephalosporins could be responsible for the amount of blaCTX-M found. The KpPi152 strain presented resistance to cephalosporins CAZ, CFC, CFM, CFZ, CPM, CRO, CRX, CTL, and CTX.

Three strains (KpPi150, KpPi153, and KpPi157) harbored the blaKPC gene (Table 1). KPC-producing bacteria are closely related to infections associated with a high level of morbidity and mortality because carbapenem antimicrobials are generally not effective against these bacteria (Arnold et al. 2011). The KpPi153 strain presented resistance to all tested carbapenems (DOR, ERT, IPM, and MPM) and KpPi150 strain to IPM and MPM. Interestingly, although the KpPi157 strain presented the blaKPC gene, it was susceptible to carbapenems. On the other hand, as described by Marschall et al. (2009) and Villegas et al. (2006), the presence of the blaKPC gene may not always result in resistance in vitro to carbapenems, resulting in a possible failure to detect this phenotype during routine workup. Therefore, an accurate investigation to detect KPC and the real resistance to carbapenems is crucial to establish control of their silent dissemination in strains that remain susceptible in vitro (Marschall et al. 2009; Villegas et al. 2006).

Occurrence of virulence genes and HM phenotype

Several virulence genes were found among the strains. All strains presented the genes fimH and ycfM, followed by mrkD (92.3%), entB (69.2%), ybtS (53.8%), and kfu (38.5%) (GenBank accession numbers MT330315, MT330318, MT330321, MT330323, MT330325, and MT330326) (Table 1). The high prevalence of fimH, mrkD, and entB genes found in this study was already expected, since several studies have shown a high correlation of these genes in K. pneumoniae strains isolated from different sources and countries (Azevedo et al. 2019; Ferreira et al. 2019; Gonçalves et al. 2017). Similar results were obtained by Kuş et al. (2017), who demonstrated that the four virulence genes most detected in K. pneumoniae strains of nosocomial infections were entB (96.2%), ycfM (86.8%), mrkD (83%), and fimH (64.2%). The strains KpPi149 and KpPi150 presented HM phenotype in addition to being MDR and XDR, respectively. The rmpA and magA genes are associated with this phenotype and some studies have shown that magA is associated to the K1 capsule-specific wzy gene (Catalán-Nájera et al. 2017). However, none of these genes were detected in both strains. This characteristic has already been reported in some Klebsiella spp., including K. pneumoniae (Lee et al. 2010; Yu et al. 2006; Mohammed and Flayyih 2018). Recently, an HM Klebsiella variicola subsp. variicola without the genes rmpA and magA was described causing primary endodontic infection and none of these genes were detected (Nakamura-Silva et al. 2020). These studies corroborate our results, indicating that other genes may be involved in the HM phenotype.

Molecular typing of K. pneumoniae strains

MLST analysis showed ten different STs (ST14, ST17, ST20, ST29, ST45, ST101, ST268, ST1800, ST3995, and ST3996) (Table 1). Two STs (ST3995 and ST3996) were described for the first time in this study, since the strains presented new alleles for the tonB gene (allele 573 for ST3995 and allele 574 for ST3996). All STs were found to belong to CG258 (Fig. 2), which is considered a high-risk CG and is well known for spreading ARGs, mainly blaKPC and virulence genes worldwide (Pitout et al., 2015). CG258 and ST11 have been described as widespread in Brazil (Gonçalves et al. 2017), including outpatient infections (Azevedo et al. 2019).

Fig. 2
figure 2

A goeBurst diagram representing the clonal relationship among STs of K. pneumoniae strains indicating clonal group 258 (CG258) (blue) and sequence types (STs) found (red)

Full size image

The new ST3996 was found in an XDR strain (KpPi156) which is a single-locus variant (SLV) of ST11 (Fig. 2). Two MDR strains (KpPi146 and KpPi153) presented ST101, which is considered an emerging clone that has been identified worldwide and has the potential to become a persistent threat to global public health (Roe et al. 2019). Moreover, the Kpi153 strain presented the blaKPC gene (Table 1). These characteristics were also found in southern Brazil and Italy (Gonçalves et al. 2017; Roe et al. 2019). Another new ST found in this study was ST3995 that was described in the strains Kpi152 and Kpi159, which is also an SLV of ST101. The appearance of strains with the same ST in different patients and sources suggests the spread of these STs, a characteristic observed in ST3995, ST101, and ST17 (Table 1).

The Kpi145 strain presented ST268/KL20, which corroborates with several other studies that demonstrate an association between ST268 and K20 in K. pneumoniae strains (Table 1) (Liu et al. 2014; Lin et al. 2015; Yan et al. 2015, 2016; Zhang et al. 2016; Guo et al. 2017; Chen et al. 2017). The KpPi148 strain (ST14) harbors the blaOXA-1-like gene and other β-lactamase-encoding genes that have already been described in this ST, such as K. pneumoniae strains harboring blaCTX-M-15 causing neonatal sepsis in Tanzania (Mshana et al. 2013), and K. pneumoniae co-producing blaOXA-48 and blaNDM carbapenemases in Dubai (Moubareck et al. 2018). Furthermore, the KpPi148 strain presented the capsular serotype K2, which is frequently associated with invasive infections when found in hv clones (Wyres et al. 2019); however, ST14 is a non-hv clone and K. pneumoniae ST14/K2 strains have been found in several other studies (Brisse et al. 2009; Harada et al. 2018; Musicha et al. 2019).

The KpPi147 presented ST45, wzi allele 101, and associated K-type 24 (Table 1). These same characteristics were found in an outbreak caused by MCR-1-producing K. pneumoniae strains (24 strains of ST45 and one of ST1112) isolated from patients in a hospital in Porto, Portugal (Mendes et al. 2018). Moreover, in a study of ESBL-producing Enterobacteriaceae causing sepsis in neonates at a tertiary hospital in Tanzania, 18 out 38 K. pneumoniae strains showed ST45, being the most common ST among the strains; however, unlike KpPi147, all K. pneumoniae ST45 strains in the African study presented the blaCTX-M-15 gene (Marando et al. 2018). ST20 has already been reported causing outbreaks in neonatal wards. Between 2012 and 2013, Jin et al. (2015) described two outbreaks in China involving MDR K. pneumoniae ST20 and ST17 harboring blaNDM and blaKPC. In comparison, our strain KpPi155 (ST20) did not show any of these genes. In a recent review of the genomic population of K. pneumoniae, Wyres et al. (2020) described eight MDR global problem K. pneumoniae clones, including ST101 and ST20, which were found in our strains (Table 1).

The strains KpPi149 and KpPi150 presented ST17, both have wzi allele 141, similar virulence genes, and HM phenotype, but only Kpi150 presented the blaKPC gene and an XDR profile (Table 1). ST17 has also been shown to carry other ARGs in southern Brazil, such as blaOXA-370 and blaCTX-M-8 (Aires et al. 2016). In addition, ST17 was also found in a CTX-M-15-producing K. pneumoniae strain in Norway, which caused an outbreak in a neonatal intensive care unit and subsequent intestinal colonization of affected children for up to 2 years (Löhr et al. 2015). Some STs found in this study were described as causing neonatal outbreaks in other countries (Mshana et al. 2013; Marando et al. 2018; Jin et al. 2015; Löhr et al. 2015). However, in the present study, none of the strains were isolated from neonates, with the average age of the patients being 57 years old and the youngest a child of 3 years old (KpPi159) (Table 1).

In the present study, the strain KpPi151 presented ST29 with wzi allele 85 and K-type 30. Moura et al. (2017) analyzed the genome sequence of an ST29 HM/hv MDR CTX-M-15 K. pneumoniae strain isolated from human infection in southeastern Brazil; however, they found the wzi allele 19 and K-type 19 for this strain. To the best of our knowledge, there are no reports of the ST1800 of K. pneumoniae ST1800 (KpPi157) in the literature, where only two strains are deposited in the K. pneumoniae MLST database (https://bigsdb.pasteur.fr/klebsiella/klebsiella.html), however, without any information about them. The wzi sequences submitted to the MLST database did not find alleles corresponding to K-type for the strains KpPi149, KpPi150, KpPi153, and KpPi155 (Table 1).

Conclusions

The results presented in this study raise worrying data related to multidrug resistance and virulence in K. pneumoniae strains isolated from inpatients in the northeast of Brazil. In addition, all STs found, including two new STs, belonged to CG258, an international high-risk CG. Therefore, this study contributes to epidemiological surveillance studies in monitoring MDR, XDR, and virulent clones pertaining to CG258 worldwide.