Deciphering the gastrointestinal carriage of Klebsiella pneumoniae

ABSTRACT

Bacterial infections pose a significant global health threat, accounting for an estimated 7.7 million deaths. Hospital outbreaks driven by multi-drug-resistant pathogens, notably Klebsiella pneumoniae (K. pneumoniae), are of grave concern. This opportunistic pathogen causes pneumonia, urinary tract infections, and bacteremia, particularly in immunocompromised individuals. The rise of hypervirulent K. pneumoniae adds complexity, as it increasingly infects healthy individuals. Recent epidemiological data suggest that asymptomatic gastrointestinal carriage serves as a reservoir for infections in the same individual and allows for host-to-host transmission via the fecal-oral route. This review focuses on K. pneumoniae’s gastrointestinal colonization, delving into epidemiological evidence, current animal models, molecular colonization mechanisms, and the protective role of the resident gut microbiota. Moreover, the review sheds light on in vivo high-throughput approaches that have been crucial for identifying K. pneumoniae factors in gut colonization. This comprehensive exploration aims to enhance our understanding of K. pneumoniae gut pathogenesis, guiding future intervention and prevention strategies.

INTRODUCTION

Carl Friedländer’s identification of bacteria in the lungs of pneumonia patients in 1882 marked a pivotal moment, attributing the disease to bacterial causation, naming the bacterium Friedländer bacillus, which was subsequently named Klebsiella pneumoniae (K. pneumoniae) after microbiologist Edwin Klebs (1–3). In the ensuing 141 years, K. pneumoniae, a gram negative, encapsulated, non-motile, and rod-shaped bacterium, has been extensively studied as it can readily colonize human mucosal surfaces and is considered an opportunistic pathogen that can cause the development of pneumonia, bacteremia, pyogenic liver abscesses, and urinary tract infections (UTIs) (Fig. 1A) (4). What has been historically referred to as “Klebsiella pneumoniae” has since been demonstrated to be a complex of very closely related Klebsiella species (5–8), known as the K. pneumoniae species complex (KpSC), which comprises clinically relevant species that include K. pneumoniae, K. quasipneumoniae subsp. quasipneumoniae and similipneumoniae, K. variicola subsp. variicola and tropica, K. quasivariicola, and K. africana. This review will focus on K. pneumoniae sensu stricto, comprising ~85% of KpSC isolates (5, 9–12).

Fig 1.

Schematic of K. pneumoniae-associated disease states and the model for route of gastrointestinal colonization and host-to-host transmission. (A) Disease states associated with the two pathotypes of K. pneumoniae. (B) Model representation of K. pneumoniae colonization of the mucosa of the gastrointestinal tract and transmission, based on the epidemiological and murine models of colonization and transmission.

Considered a significant contributor to nosocomial infections, K. pneumoniae poses a mounting challenge in clinical settings due to its increased resistance to frontline antibiotics commonly used to treat bacterial infections (13, 14). Studies suggest that K. pneumoniae causes approximately 12% of pneumonia cases, 2%–6% of UTIs, and has a high mortality rate of 27%–37% in bacteremia cases, which has led to a surge in antimicrobial resistance (AMR) (15–20). K. pneumoniae AMR is partly owed to an increased capacity for acquiring and retaining AMR genes (ARGs) from bacteria spanning different families, orders, classes, and phyla (21, 22). In addition to acquiring ARGs, K. pneumoniae acts as a crucial reservoir of ARGs, facilitating the dissemination of resistance to other pathogenic bacteria (21–24). Recognizing the growing threat of AMR and the persistent rise in hospital-acquired infections (HAIs), the CDC and WHO have classified K. pneumoniae as a critical pathogen, underscoring the urgent need to develop innovative treatment strategies (25, 26). Furthermore, the Infectious Disease Society of America (IDSA) has included K. pneumoniae in a list of pathogenic bacteria that have rapid antibiotic resistance development: ESKAPE (Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa, and Enterobacter spp.) (27).

It is important to note that while K. pneumoniae disease states often involve the lungs, liver, and blood, the initial colonization often occurs in the gastrointestinal (GI) tract (28, 29). GI colonization is predominantly asymptomatic, but recent studies suggest that the strain responsible for disease manifestations in an individual is frequently present in the same patient’s GI tract (7, 29–33). This duality positions K. pneumoniae as a potential “pathobiont” capable of both asymptomatic carriage and disease state. The covert nature of gut colonization intensifies the threat, rendering outbreaks that are challenging to track, as asymptomatic carriers may disseminate the pathogen across considerable spatial or temporal distances (7, 29, 30, 34–38). While extensive research has delved into K. pneumoniae virulence at extra-intestinal sites, the initial GI tract colonization has been relatively overlooked. This review aims to consolidate existing knowledge on K. pneumoniae GI colonization, providing a foundation to enhance future investigations.

EPIDEMIOLOGICAL INSIGHTS INTO K. PNEUMONIAE GI COLONIZATION

Pioneering work by Selden et al. in 1971 during a multi-drug-resistant (MDR) K. pneumoniae outbreak at a Denver hospital unveiled the silent gut reservoir of this pathogen. The study revealed that 25% of the ICU patient population and several hospital staff were positive for K. pneumoniae gut colonization, with approximately 4% identified as oropharyngeal carriers. Notably, a significant proportion of carriers progressed to infection by the same K. pneumoniae serotype, underscoring the pivotal role of GI carriage as an intermediate step in developing K. pneumoniae-associated diseases and as a reservoir for host-to-host transmission (Fig. 1B) (39).

Recent years have witnessed outbreaks with alarming frequency, particularly in ICUs, linked to MDR K. pneumoniae (37, 40, 41). A 2011 outbreak at the National Institute of Health Clinical Center showcased the limitations of conventional epidemiology in elucidating K. pneumoniae colonization and transmission dynamics (30). The genome sequencing of patient isolates revealed that the outbreak originated from a single parental strain, with transmission tracked via single-nucleotide variants. Three independent transmission events were identified from the initial infected patient, with many patients harboring K. pneumoniae in the GI tract, surviving on hospital equipment, and potentially colonizing healthcare workers (30). Follow-up studies have shown that the GI tract is a reservoir for dissemination and host-to-host transmission and foster antibiotic resistance development (7, 23, 29, 37). Another study demonstrated that even though treatment of carbapenem-resistant K. pneumoniae (CRKP) infection with colistin led to clearance from extra-intestinal sites, it resulted in colistin resistance and persistence in the GI tract, highlighting the gut’s significance as a reservoir (42).

While many studies have focused on MDR-K. pneumoniae outbreaks in a hospital setting, a 2017 study provided a comprehensive view of K. pneumoniae carriage in the GI tract (7). Screening of patients admitted to an ICU revealed that ~6% of patients admitted directly from the community and ~19% from other healthcare facilities were colonized with K. pneumoniae in the GI tract. This study, conducted on a larger cohort without an outbreak context, showcased a broader diversity of K. pneumoniae strains in patients’ GI tracts. Consistent with earlier findings (30, 39), colonizing strains often matched those causing subsequent pathology. Similarly, a longitudinal study in a hospital setting by Martin et al. (32) observed that 23% of patients had K. pneumoniae in their GI tracts. Furthermore, their results showed a significant association between gut colonization and subsequent infection with the same isolate. Together, these studies provide strong evidence of GI colonization preceding infection.

Numerous risk factors contribute to K. pneumoniae carriage, including the use of medications such as proton pump inhibitors, non-steroidal anti-inflammatory drugs, recent antibiotic administration, and other factors such as travel to Asia, and having Crohn’s disease/ulcerative colitis (38). In community settings, colonized individuals can transmit K. pneumoniae to close family members (familial transmission) and household pets. In low-income countries, the colonization dynamics of K. pneumoniae are further influenced by rapid urbanization and seasonality, indicating that environmental factors drive persistence and transmission (36, 43, 44). These epidemiological investigations underscore the ability of K. pneumoniae to colonize both healthy and immunocompromised populations asymptomatically. GI carriage rates in community settings vary globally, ranging from approximately 19%–88% in Asia, 5%–25% in Western countries, and 23%–70% in hospitalized patients, with antibiotic usage and length of in-patient stay also contributing to an increase in K. pneumoniae GI carriage (32, 45–48). Considering the elevated carriage rate in hospitalized patients, persistence in a community setting, and MDR status, a comprehensive understanding of the initial stages of K. pneumoniae colonization is imperative. This knowledge is essential for developing prevention therapies and clinical strategies to effectively manage colonization before it progresses to pathological infections.

Accordingly, GI decolonization strategies using selective digestive tract decontamination (SDD) using antibiotics and fecal microbiota transfer (FMT) against K. pneumoniae have become more appealing to mitigate the development of the disease states. SDD using antibiotics has been successfully used to reduce K. pneumoniae burden in the gut. However, SDD treatment can also result in secondary resistance to antibiotics being used for decolonization (49, 50). FMT is an attractive alternative strategy that does not involve antibiotic treatment and has been successfully used to reduce Clostridium difficile burden in the gut (51). FMT treatment reduced K. pneumoniae carriage in most patients (>70%), with significant changes in microbial diversity and corresponding changes in short-chain fatty acids (SCFAs) and bile acids (52–54). Thus, FMT may provide a route in the identification of microbiota members that selectively reduce K. pneumoniae colonization.

DESCRIPTION OF K. PNEUMONIAE PATHOTYPES

K. pneumoniae, known for its adaptability in acquiring genetic material, manifests into two distinct pathotypes: classical K. pneumoniae (cKp) and hypervirulent K. pneumoniae (hvKp). The cKp pathotype is predominantly associated with HAIs and exhibits MDR. By contrast, the hvKp pathotype possesses additional virulence traits, enabling infections in immunocompetent individuals within a community setting. Traditionally, these two pathotypes were considered separate because they exhibited distinct genetic backgrounds (55, 56). Although not common, a noteworthy development involves the recent emergence of convergent clones that have acquired genetic material, rendering these strains both MDR and capable of causing severe infections in otherwise healthy individuals, thus posing a substantial public health threat (57–59).

cKp isolates under antibiotic pressure undergo continuous evolution, accumulating antibiotic resistance genes (ARGs) through plasmid acquisition, genetic element integration, de novo mutations, or a combination of these mechanisms. Over 50 antibiotic-resistant plasmids and hundreds of ARGs have been identified in the cKp genome, conferring resistance against major antibiotic classes such as cephalosporins, aminoglycosides, fluoroquinolones, and carbapenems (60, 61). Concerningly, K. pneumoniae AMR rates have steadily increased, with MDR cKp becoming endemic in many countries. Besides AMR, the cKp strain can belong to any of the 134 distinct capsular serotypes (K type) that produce primarily the siderophore enterobactin, with some strains also producing the yersiniabactin siderophore (48, 62–66). Among the classical strains, two sequence types (ST) are considered high risk (epidemic or problem clones) (67): ST45, which encodes extended-spectrum beta-lactamases (ESBLs), enzymes that confer resistance against third-generation cephalosporins (68), and ST258, which encodes carbapenemases (69). Pathologically, cKp is most commonly associated with UTIs, pneumonia, and bacteremia (70).

While the absence of distinct genotypic or phenotypic markers complicates the differentiation of cKp and hvKp strains, several distinctive features define hvKp. Compared to cKp, hvKp strains have lower serotype diversity, with most K1 and K2 serotype strains displaying a hypermucoviscous (HMV) phenotype (71–74). hvKp strains produce additional siderophores, including yersiniabactin, salmochelin, and aerobactin, but infrequently produce ESBLs or carbapenemases (55, 56, 70, 75–78). The hvKp phenotype is primarily driven by a 100- to 200-kb virulence plasmid encoding the additional siderophores and genes that drive the HMV phenotype (79–81). Despite the genetic plasticity and high plasmid retention contributing significantly to the expanding repertoire of AMR in cKp (21, 22), conventional hvKp strains remarkably remain susceptible to currently available antimicrobials. The scarcity of ARGs in hvKp is proposed to result from increased capsule production and the HMV phenotype, likely acting as a barrier for DNA uptake (82, 83). In addition, the activity of CRISPR/Cas systems prevents the integration of any acquired DNA (84). The presence of a large virulence plasmid may further impose a fitness cost because of a higher metabolic burden, resulting in reduced uptake of AMR genes (85–87). Convergent strains that are phenotypically both cKp and hvKp are primarily derived by cKp isolates acquiring hybrid plasmids containing ARG and hypervirulence genes. ST307 lineage cKp, which generally carry ARGs for carbapenems and newer-generation cephalosporins, were recently observed also to carry genes associated with hypervirulence, highlighting this convergent phenotypic phenomenon (88, 89).

hvKp is often associated with a broader spectrum of infections than its classical counterpart, including pyogenic liver and splenic abscesses, endophthalmitis, and meningitis (Fig. 1A) (75, 76). A distinctive hallmark of hvKp lies in its ability to metastatically spread to multiple infection sites in immunocompetent hosts within a community setting, an uncommon trait among enteric gram negative bacilli (75, 76). Both K. pneumoniae pathotypes share the GI tract as their initial colonization site (7, 29, 30, 36, 39, 90), emphasizing the significance of studying GI colonization to elucidate the commonalities and distinctions between pathotypes and develop strategies to mitigate pathogen spread.

ANIMAL MODELS TO STUDY K. PNEUMONIAE GUT COLONIZATION

While in vitro studies have significantly contributed to our comprehension of K. pneumoniae pathogenesis, their limitations in replicating the dynamic conditions within the host necessitate in vivo investigations. Existing animal studies, particularly those utilizing murine models, have predominantly focused on various disease states induced by K. pneumoniae, often overlooking the crucial aspect of GI colonization (91–94). Recent epidemiological studies (30, 32, 39) recognizing the GI tract as a prominent reservoir for K. pneumoniae have spurred the development of murine models that delve into the molecular mechanisms underlying the interactions between K. pneumoniae and the host. These models intricately explore the role of the microbiota, antibiotic treatment, and host immune responses in the context of colonization (Table 1).

TABLE 1.

Animal models of K. pneumoniae gastrointestinal colonization

Mode of inoculation	Treatment	Genetic background	Pathotypes tested	Experimental conditions
Intragastric	Oral antibiotic	Mus musculus; CFW1, C57BL/6, OF1	cKp	Colonization density, bacterial factors, mutagenesis screen, histopathology, fluorescent in situ hybridization (95–99)
Intragastric	Sodium bicarbonate	Mus musculus; C57BL/6	cKp and hvKp	Histopathology, 16S analysis, colonization density, host response, bacterial factors (100)
Intragastric	None	Mus musculus; BALB/c	hvKp	Virulence, colonization density, bacterial factors (101)
Oral feeding	None	Mus musculus; C57BL/6	cKp and hvKp	Histopathology, 16S analysis, colonization density, virulence, bacterial factors, transmission (102–104)
Intragastric	None	Mus musculus; C57BL/6 (germ-free)	cKp	Colonization density, microbiome, phage therapy (105, 106)
Intragastric	Intraperitoneal/ subcutaneous antibiotic	Mus musculus; CF1, C57BL/6	cKp	Colonization density, mutagenesis screen, bacterial factors (106, 107)
Intragastric	None	Gallus domesticus; PA12 (germ-free)	cKp	Colonization density, bacterial factors (108)

The extensively employed model using intragastric inoculation of antibiotic-treated mice has been pivotal in unraveling the intricate dynamics of pathogen-host interactions (106, 109–120). This antibiotic treatment model holds particular significance in portraying K. pneumoniae interactions within a dysbiotic gut, closely mimicking human infections within a hospital setting. It serves as a valuable tool for understanding the impact of antibiotics on host-to-host transmission, shedding light on the conditions prevalent in a hospital setting (30, 121). Despite its merits, the antibiotic treatment model leads to a reduction in gut microbial diversity, which limits the exploration of interactions between K. pneumoniae and the native gut microbiota. Notably, K. pneumoniae colonizes individuals such as hospital staff and others in a community setting in the absence of antibiotic treatment (29, 34–36, 106, 110, 119).

Recent studies (100, 102) have addressed the limitations by developing murine models to study K. pneumoniae’s GI colonization in the presence of an intact microbiome. Calderon-Gonzales et al. (100) introduced the quenching model, neutralizing stomach acid via oral gavage of sodium bicarbonate before K. pneumoniae inoculation without antibiotics. Young et al. (102) utilized the oral feeding method with a pipette, mimicking the natural colonization route to establish stable gut colonization. Both models facilitate a nuanced understanding of the role of the microbiome in resisting incoming pathogens, which allows for ascertaining the intricacies of pathogen-microbiota interactions. The quenching model coupled with the standard inoculation technique of oral gavage offers accessibility for laboratories familiar with the method. The oral feeding model has been instrumental in elucidating the transmission dynamics of K. pneumoniae both with and without antibiotics and revealed increased transmission efficiency following antibiotic treatment. Collectively, these animal models align closely with epidemiological findings and offer a robust foundation for conducting mechanistic studies involving K. pneumoniae interactions within the GI tract.

KLEBSIELLA PNEUMONIAE INTERACTIONS IN THE GI TRACT

The human GI tract hosts a diverse community of microorganisms comprising bacteria, fungi, protozoa, viruses, and archaea. This intricate microbial ecosystem maintains a symbiotic relationship with the host and protects against the colonization of incoming pathogens, known as colonization resistance (CR) (122). The orchestration of CR involves the collaborative efforts of resident microbiota and host defenses (123–126).

CR employs various mechanisms, broadly categorized as direct or indirect antagonistic measures. Direct inhibitory mechanisms including nutritional and spatial competition (127, 128), contact-dependent inhibition (129), microbiota-derived antimicrobial peptide (AMP) release (130), and inhibitory metabolites, including secondary bile acids (123) and SCFAs (131). Indirect inhibition involves the mucosal barrier preventing access of pathogenic bacteria to the underlying gut epithelial layer, induction of intestinal hypoxia to limit the expansion of facultative anaerobes, and the release of host-derived antimicrobial peptides and proteins, such as Lipocalin 2 (132–134). Other contributions to CR come from the activation of cytokines such as IL-22 that modulate intestinal barrier function (135), and the host’s adaptive immune response, including homeostatic IgA (sIgA) that can cross-react with antigens present on enteric pathogens (136).

The delicate balance of the gut microbiota is susceptible to disruption or alteration through various factors such as diet, immune suppressants, proton pump inhibitors, or antibiotics, and the imbalance can render the host more vulnerable to pathogenic bacteria, including K. pneumoniae (125, 137–139). Extensive evidence has demonstrated that the association between antibiotic treatment in a hospital setting and the expansion of K. pneumoniae in the gut subsequently leads to disease onset (70, 75). However, the presence of K. pneumoniae in the GI tract of individuals in a community setting suggests that it is a pathobiont capable of overcoming microbiota-mediated CR through inherent mechanisms (45, 46, 48, 76). This resilience underscores the complex interplay between K. pneumoniae and the host, emphasizing the need for a nuanced understanding of the factors influencing gut microbiota dynamics in the context of pathogenic bacterial colonization.

Freter et al. (140) postulated the nutritional niche theory, rooted in the concept of direct competition, which dictates that the ability of a pathogen to thrive depends on the availability of specific nutrients. Generally, species compete with other closely related species, as they exhibit overlapping substrate ranges. Using this principle, Osbelt et al. demonstrated that K. pneumoniae is outcompeted by Klebsiella oxytoca through augmented metabolic capacity (beta-glucoside metabolism) conferred by the casRAB operon (Fig. 2A) (141).

Fig 2.

Mechanisms of colonization resistance (CR) against K. pneumoniae and the ability of K. pneumoniae to overcome CR. (A) Native microbiome members such as K. oxytoca that have an expanded metabolic capacity and occupy the same niche as K. pneumoniae can provide CR against K. pneumoniae colonization. The resident gut microbiota members can also limit the colonization of K. pneumoniae through modulation of the host immune response through a commensal consortium of Bacteroidetes. (B) K. pneumoniae can overcome CR through the metabolism of alternative nutrient sources that allow it to bypass microbiota-mediated nutritional limitation and colonize the gut. The use of the Type 6 secretion system (T6SS) allows K. pneumoniae to directly kill commensals that occupy the same spatial niche. An indirect mechanism of inhibition is through the host release of antimicrobial peptides that can limit pathogen growth, which K. pneumoniae overcomes through the SAP transporter.

Alternatively, invading microbes may bypass nutrient competition using less “preferred” nutrients that are energetically less efficient (142–145). K. pneumoniae possesses a diverse array of genes that encode for carbohydrate metabolism, which suggests the potential to overcome CR (146). Hudson et al. (104) hypothesized that, in the gut environment, K. pneumoniae utilizes alternative nutrients to navigate intense competition and bypass CR. They demonstrated that the ability to metabolize fucose, liberated from mucins by enzymes produced by Bacteroides (147–149), facilitates K. pneumoniae GI colonization (150). Ethanolamine (EA), a byproduct of cell membranes, can be metabolized by enteric pathogens in the GI tract to overcome CR and modulate virulence phenotypes. In an ICU setting, a dominant CRKP ST11 isolate exhibited an enhanced ability to metabolize EA, translating to successful colonization in the murine GI tract, concomitant with reduced virulence. This highlights the role of EA metabolism as a pivotal mechanism enabling K. pneumoniae to overcome CR, and serving as a selective force that facilitates the dissemination of MDR clones (151). Cellulose, a major component of plant cell walls, is neither digested nor broken down by the host but can be degraded by gut commensals to release cellobiose (152). The putative cellobiose phosphotransferase system (PTS) of the hvKp strain, NTUH-K2044, can transport this sugar (153). In vivo competition assay experiments have identified that the cellobiose PTS system is critical for K. pneumoniae gut colonization in the presence of the native microbiota (153). Another alternative carbon source that hosts and commensals use is trehalose, a non-reducing disaccharide that is part of a normal diet. Trehalose can be broken down into glucose and incorporated into the glycolytic pathway (154). An NTUH-2044 mutant lacking trehalose-6-phosphate hydrolase (treC) had an impaired ability to form biofilms and colonized the murine GI tract poorly (155). These results indicate that the ability to metabolize cellobiose and trehalose are crucial metabolic determinants, allowing K. pneumoniae to bypass nutritional competition in the GI tract. As a urease-positive bacterium, K. pneumoniae can hydrolyze urea, which is secreted into the GI tract by the host as a byproduct of the urea cycle (156). Urea hydrolysis produces ammonia, which is competed for by the host and the resident microbiota including K. pneumoniae. A urease deletion mutant was unable to colonize the gut, implying that urea serves as an important nitrogen source in the gut (Fig. 2B) (111).

Type VI secretion systems (T6SS), present in some gram negative pathogens, is a contact-dependent inhibition system with antibacterial activity (157, 158). Both commensals and pathogens encode T6SSs, suggesting that commensals deploy it as a mechanism of CR, whereas, conversely, pathogens deploy it to overcome CR. In the case of K. pneumoniae, a demonstrated reliance on the T6SS for gut colonization underscores its significance in niche occupancy within the GI tract (Fig. 2B) (100, 159, 160).

An indirect mechanism of CR against K. pneumoniae colonization is through the modulation of host immune responses by the gut commensal Bacteroidetes. This protective mechanism against K. pneumoniae colonization hinges on the Bacteroidetes commensal colonization factors (CCF), which foster the association of Bacteroidetes with the mucosa. This association triggers IL-36 signaling and subsequent macrophage recruitment, collectively acting as a defense against K. pneumoniae colonization (Fig. 2A) (113). Another indirect mechanism of CR is the release of host-derived AMPs. However, bacterial pathogens, including K. pneumoniae, can resist host AMPs through the Sap (sensitivity to antimicrobial peptides) transporter (161–163), thus implicating the Sap transporter as a critical factor contributing to the ability of K. pneumoniae to overcome host-mediated CR (Fig. 2B).

K. pneumoniae colonization of the permissive environment of the neonatal gut, which consists of predominantly aerotolerant bacteria (164), can lead to subsequent development of colitis, with a potential role for polyketide synthase (pks) island that encodes colibactin, a genotoxin capable of inducing DNA damage (116, 118). However, these studies used immunologically deficient (Il10-/-, T-bet-/- and Rag2-/-) gnotobiotic mice. Thus, further investigation using a more complex microbiome is required to fully elucidate the factors and interactions driving K. pneumoniae-associated gut inflammation (116, 118). Ultimately, an intricate network of interactions and evolutionarily honed molecular mechanisms determine the ability to colonize and persist. Understanding the underlying mechanisms of K. pneumoniae interactions in the gut and the nutrients it utilizes to overcome CR is crucial in preventing and containing this dangerous pathobiont.

KNOWN CONTRIBUTORS TO K. PNEUMONIAE GI COLONIZATION AND HOST-TO-HOST TRANSMISSION

Capsule

Reverse genetics have been used to identify virulence and colonization determinants of K. pneumoniae. One of the well-studied virulence factors of K. pneumoniae is its capsular polysaccharide (CPS). CPS provides K. pneumoniae with protection against complement-mediated clearance, AMPs, and even reduces the inflammatory response via NOD1 activation (70, 165, 166). In initial studies, the role of CPS in gut colonization appeared contradictory (95, 167, 168). Those studies were performed with mice treated with antibiotics, so they likely masked the effect of K. pneumoniae factors. Two capsule-deficient mutants (wcaJ and manC) were found to have colonization defects in a GI model with an intact microbiota (no antibiotic treatment) (102). Furthermore, a mgrB mutant, which results in colistin resistance, caused a pleiotropic effect of reduced CPS production, also resulting in a defect in GI colonization in the presence of gut microbiota (103). Reduced CPS levels correlate with increased binding to mucin proteins that form the mucus layer lining the GI tract (103). Thus, the robust GI colonization of K. pneumoniae in the presence of gut microbiota is likely due to its interactions with the gut mucus layer mediated through CPS, which is dispensable when microbial diversity is reduced under antibiotic pressure.

Type 1 and 3 fimbriae

Most K. pneumoniae strains produce both type 1 and type 3 fimbriae (70). Type 1 fimbriae are thin, thread-like structures on the bacterial cell surface and are commonly associated with adherence to mucosal surfaces (169). Type 3 fimbriae have a helix-like structure and have been shown to bind host extracellular matrix components such as type IV and V collagens (170). There have been conflicting results regarding the role of type 1 and type 3 fimbriae in GI colonization (96, 112, 171). In a model with antibiotic treatment, both the type 1 and 3 fimbriae were dispensable for GI colonization with a UTI isolate (C3091) (96, 171). By contrast, Khater et al. observed that both type 1 and 3 fimbriae were essential for GI colonization with a wound isolate (LM21) (112). Furthermore, in a murine model with an intact microbiome, type 1 fimbriae of the strain KPPR1 (lung isolate) were observed to be dispensable for GI colonization (102). The type 1 fimbriae loci (fim) associated with UTI and lung isolates were found to be in the “off” position, indicating their inactivity in the GI tract. These observed differences could be attributed to experimental design or the genetic plasticity of K. pneumoniae, highlighting the importance of testing virulence determinants under different conditions to better understand their importance.

Antibiotic treatment

Several studies have highlighted the effect of antibiotics on K. pneumoniae gut colonization, where growth suppression of anaerobes or Bacteroides promoted K. pneumoniae colonization in the gut (106, 115, 172) and suggested that specific members of the microbiome are involved in actively providing CR against K. pneumoniae (173). Epidemiological studies have demonstrated a correlation between antibiotic treatment, an increase in bacteremia, and an increase in host-to-host transmission (174, 175). Young et al. (102) demonstrated that antibiotic treatment in mice mediated a reduction in microbial diversity and correlated with a bloom of K. pneumoniae in the gut. This bloom, called the “supershedder” state (176), allows K. pneumoniae to transmit at a higher frequency. However, unlike other enteric pathogens (177, 178), this supershedder state is entirely dependent on the continuation of antibiotic treatment, as removal of antibiotics leads to resetting of the microbiome and reduction in K. pneumoniae shedding (102). It has been demonstrated that K. pneumoniae colonized mice with depleted microbiota due to antibiotic treatment show a decrease in colonization after discontinuing antibiotics and receiving FMT from healthy donors. This suggests that restoring the microbiome after antibiotic treatment could be a meaningful way to control K. pneumoniae expansion (98).

Factors affording gut survival

Bacterial factors that confer an advantage in surviving the harsh environment of the GI tract would likely be necessary for gut colonization. An efflux pump encoded by eefABC is induced by acidic and hyperosmotic conditions in the GI tract and is required for K. pneumoniae GI colonization (179). Furthermore, the cad (lysine decarboxylation) and tdc (threonine and serine metabolism) operons of the hvKp isolate NTUH-K2044 were required for resistance to inorganic acids and bile salts and consequently identified to be important for gut colonization in the presence of native gut microbiota (101, 179). Similarly, OxyR, known to protect against oxidative stress, was required for robust K. pneumoniae gut colonization and survival in bile salts and acidic environments, both conditions that are prevalent in the GI tract (180). Interestingly, the ter operon, found in certain K. pneumoniae isolates, was identified to be an important gut colonization determinant only in the presence of microbiota species that produce SCFAs. However, the effect is not due to changes in K. pneumoniae intracellular acidification but likely due to some unknown mechanism (181). Collectively, these studies highlight the different strategies that K. pneumoniae employs to colonize and persist in the GI tract and provide a solid foundation for further elucidating the mechanisms and interactions mediating K. pneumoniae GI colonization.

MUTAGENESIS SCREENS TO IDENTIFY K. PNEUMONIAE GUT DETERMINANTS

Forward genetic screens, particularly transposon insertion sequencing (TIS), have revolutionized genetic screens by combining transposon mutagenesis with next-generation sequencing to determine bacterial factors that contribute to fitness and or essentiality under a condition of interest. TIS encompasses transposon sequencing (Tn-seq) (182), insertion sequencing (INSeq) (183), transposon-directed insertion site sequencing (TraDIS) (184), and high-throughput insertion tracking by deep sequencing (HITS) (185). Herein, we will use TIS to refer to a technique that involves the construction of a saturated library of random mutants by inserting a transposon element, followed by sequencing the chromosomal regions flanking the transposon. This approach allows for the determination of genes that do not have any insertions and, thus, identifying potentially essential genomic regions. The mutant library can also be subjected to a selective parameter to identify critical genomic regions for surviving and growing in the selective condition. Comparing the frequency of insertions in the initial population versus the post-selection population allows for the identification of the genomic regions important for fitness in the test conditions. (Refer to Cain et al. (186) for an in-depth review of TIS.)

TIS screens have been pivotal in unraveling the genetic determinants that influence K. pneumoniae pathogenesis. Applied to multiple K. pneumoniae strains using in vivo selection in murine and other host models (187), TIS screens have identified factors critical for bacteremia (188), fitness in the lung (189), and those that mediate neutrophil interaction during lung infections (190). Furthermore, Signature-tagged Mutagenesis (STM) (191), an earlier insertional mutation technique to identify negative and positive mutations after selection, has been applied to K. pneumoniae (192–194). Maroncle et al. used GI colonization and an in vitro cell culture model to identify factors important for cellular adhesion and GI colonization (192). They identified 16 genes important for cell adhesion, 12 for GI colonization, and one putative adhesin associated with both (192). They identified the fucose operon as critical for adhesion, which was recently observed as important for K. pneumoniae gut colonization (104). Struve et al. also examined an STM library using the murine GI tract and UTI model as a screening parameter (193). They identified genes involved in core and O-antigen lipopolysaccharide (LPS) biosynthesis as well as genes related to cell-membrane or cell-surface structures, such as phospholipid biosynthesis and outer membrane proteins (193). Interestingly, there was no overlap in the genes identified between the two STM screens. This could either be due to the different strains used, testing of a non-saturated STM library, or both. Using a CRKP ST258 strain, Jung et al. created and tested a TIS library under in vitro aerobic and anaerobic conditions and a murine model of GI colonization (195). They created a saturated library of ~150,000 unique transposon mutants that covered 99.9% coding sequences (CDS) and identified 14% of the genes to be essential. They identified 35 loci (genes and intergenic regions) as being critical for robust GI colonization, and notably, identified loci with increased fitness in the GI tract. Among the 35 loci, many included protein transport/folding and energy metabolism. The loci with enhanced fitness included genes involved in maltose metabolism, suggesting a fitness cost associated with this pathway in gut colonization during antibiotic treatment (195).

A recent study used TIS on three different clinical isolates of K. pneumoniae (107). The study found that the three strains had 487 genes in common that were necessary for growth in lysogeny broth culture. However, only 27 genes that overlapped between the three strains were necessary for GI colonization (107). Each of the three isolates had strain-specific genes that were required for colonization. As K. pneumoniae strains typically have genomes of ~5,500 genes, consisting of ~2,000 core genes and ~3,500 genes drawn from the accessory genome comprised of ~30,000 genes, it is possible that the lack of overlap in shared genes contributing to GI colonization would be attributable to a diverse array of accessory genes (5). Most genes found to be required for GI colonization by only one strain were also present in the other strains (107). This strongly suggests that a broad set of shared genes that are differentially employed contribute to K. pneumoniae GI colonization, possibly due to alternative regulation or interaction with other molecular systems/pathways derived from accessory genes (5, 196).

Table 2 lists the key putative gut-specific loci identified in each of these studies and also summarizes the factors that each study focused on but does not represent every loci identified in each screen. All four studies discussed here used different, albeit related, methods for mutant library generation, testing conditions, and analysis. Therefore, they are not necessarily directly comparable. Despite these differences, there is an observable trend in the combined results that show many of the genes are related to metabolism, either contributing directly to metabolite transport, catabolism, or regulation (Table 2) (107, 192, 193, 195). It has been predicted that approximately 37% of the K. pneumoniae pan-genome (core plus accessory genes) is related to metabolic pathways, with another 13% related to membrane transport of metabolites (5, 197). When combined with the results of these genetic screens, it appears that the metabolic flexibility of K. pneumoniae is a crucial driver of GI colonization. An important factor in considering these data is that these studies were performed using antibiotic-treated mice, significantly affecting the CR. While metabolic flexibility would arguably become even more critical when K. pneumoniae competes against native microflora for nutrients (198–201), the ablation of the microbiome minimizes the potential role of various toxins and microbiota-mediated metabolites that may be employed by the microbiota against K. pneumoniae (129, 157–159, 202–207). Overall, while these existing screens provide valuable insight into the genetic dynamics of K. pneumoniae GI colonization of patients who may already be on antibiotic therapies, a gap exists in our understanding of specific factors contributing to K. pneumoniae colonization of a healthy, unperturbed gut.

TABLE 2.

Genes contributing to K. pneumoniae GI colonization^b,c

Gene	Function/pathway	Strain	Technique	Reference
aceE	Pyruvate dehydrogenase E1 component	CRE-166, KPN46, Z4160	INseq	(107)
ackAa	Acetate kinase	MH258	INseq	(195)
acrA	Multidrug efflux pump subunit	CRE-166, KPN46, Z4160	INseq	(107)
adhE	Aldehyde-alcohol dehydrogenase	CRE-166, KPN46, Z4160	INseq	(107)
arcBa	Aerobic respiration control sensor protein	MH258, CRE-166, KPN46, Z4160, C3091	INseq, STM	(107, 193, 195)
bamBa	Outer membrane beta-barrel assembly protein	MH258	INseq	(195)
bamE	Outer membrane beta-barrel assembly protein	MH258	INseq	(195)
bglY	Beta-galactosidase	CRE-166, KPN46, Z4160	INseq	(107)
carA	Carbamoyl phosphate synthase small subunit	KPN46, Z4160	INseq	(107)
carB	Carbamoyl phosphate synthase large subunit	CRE-166, Z4160	INseq	(107)
csrDa	Putative lipoprotein	MH258	INseq	(195)
cvpA	Colicin V production protein	CRE-166, KPN46, Z4160	INseq	(107)
cyaAa	Adenylate cyclase	MH258	INseq	(195)
cydA	Cytochrome bd-I ubiquinol oxidase subunit 1	CRE-166, KPN46, Z4160	INseq	(107)
cyoA	Cytochrome O ubiquinol oxidase subunit II	MH258	INseq	(195)
cyoB	Cytochrome O ubiquinol oxidase subunit I	MH258	INseq	(195)
fnr	Fumarate and nitrate reduction regulatory protein	CRE-166, KPN46, Z4160	INseq	(107)
focA	Formate transporter	CRE-166, KPN46, Z4160	INseq	(107)
gatRa	Galactitol utilization operon repressor	MH258	INseq	(195)
glgP	Alpha-glucan phosphorylase	LM21	STM	(192)
glnA	Glutamine synthetase	CRE-166, KPN46, Z4160	INseq	(107)
gltB	Glutamate synthase (NADPH) large chain	MH258	INseq	(195)
gshAa	Glutamate-cysteine ligase	MH258	INseq	(195)
gshBa	Glutathione synthetase	MH258	INseq	(195)
gvR	Glycine metabolism regulator	LM21	STM	(192)
hemN	Oxygen-independent coproporphyringogen III oxidase	MH258	INseq	(195)
hrcU	Harpin type III secretion system	LM21	STM	(192)
hscAa	Chaperone protein	MH258	INseq	(195)
hscBa	Chaperone protein	MH258	INseq	(195)
hupA	DNA folding	C3091	STM	(193)
lacY	Lactose permease	MH258	INseq	(195)
miaA	tRNA dimethylallyltransferase	CRE-166, KPN46, Z4160	INseq	(107)
mtlD	Mannitol-1-phosphate 5-dehydrogenase	CRE-166, KPN46, Z4160	INseq	(107)
ntrC	Nitrogen metabolism regulator	LM21	STM	(192)
nuoM	NADH-ubiquinone oxidoreductase chain M	MH258	INseq	(195)
ompA	Outer membrane protein	C3091	STM	(193)
ompCa	Outer membrane porin C	MH258, CRE-166, KPN46, Z4160	INseq	(107, 195)
pal	Peptidoglycan-associated lipoprotein	CRE-166, KPN46, Z4160	INseq	(107)
pflA	Pyruvate formate-lyase 1-activating enzyme	CRE-166, KPN46, Z4160	INseq	(107)
pflB	Formate acetyltransferase 1	CRE-166, KPN46, Z4160	INseq	(107)
pgi	Glucose-6-phosphate isomerase	CRE-166, KPN46, Z4160	INseq	(107)
pheC	Cyclohexadienyl dehydratase	LM21	STM	(192)
plsX	Fatty acid and phospholipid synthesis	C3091	STM	(193)
ppx/gppA	Guanosine-5′-triphosphate, 3′-diphosphate pyrophosphatase	MH258	INseq	(195)
Ptaa	BioD-like N-terminal domain/phosphate acetyltransferase	MH258	INseq	(195)
ptsIa	Phosphoenolpyruvate-protein phosphotransferase	MH258, CRE-166, KPN46, Z4160	INseq	(107, 195)
purC	Phosphoribosylaminoimidazole-succinocarboxamide synthase	CRE-166, KPN46, Z4160	INseq	(107)
purH	Bifunctional purine biosynthesis protein	CRE-166, KPN46, Z4160	INseq	(107)
pyg	Glycogen phosphorylase	MH258	INseq	(195)
pykF	Pyruvate kinase I	CRE-166, KPN46, Z4160	INseq	(107)
rimO	Ribosomal protein S12p Asp88 methylthiotransferase	MH258	INseq	(195)
setA	Sugar efflux transporter A	CRE-166, KPN46, Z4160	INseq	(107)
surA	Folding of outer membrane proteins	C3091	STM	(193)
tamA	Outer membrane component of TAM transport system	MH258	INseq	(195)
tamB	Inner membrane component of TAM transport system	MH258	INseq	(195)
tatA	Sec-independent protein translocase protein	CRE-166, KPN46, Z4160	INseq	(107)
tatC	Sec-independent protein translocase protein	CRE-166, KPN46, Z4160	INseq	(107)
tolA	Tol-Pal system protein	CRE-166, KPN46, Z4160	INseq	(107)
tufA	Protein synthesis—elongation factor	C3091	STM	(193)
typAa	GTP-binding protein	MH258	INseq	(195)
uvrA	Excinuclease ABC subunit A	MH258	INseq	(195)
waaE	LPS core synthesis	C3091	STM	(193)
waaL	LPS core synthesis	C3091	STM	(193)
wbbO	O-antigen synthesis	C3091	STM	(193)
xylA	Xylose isomerase	MH258	INseq	(195)
yeiE	HTH-type transcriptional activator	CRE-166, KPN46, Z4160	INseq	(107)
ygjA	DedA family inner membrane protein	MH258	INseq	(195)
ygjD	O-sialoglycoprotein endo-peptidase	LM21	STM	(192)

^a Denotes a gene that was also found to significantly impact growth under in vitro conditions.

^b Summary of genes identified in previous Tn-seq and STM screens as contributing to K. pneumoniae GI colonization.

^c CRE-166: carbapenem-resistant lung isolate, ST258. KPN46: non-epidemic, antibiotic susceptible blood isolates, ST433. Z4160: ESBL-producing blood isolates, ST45. LM21: cutaneous wound isolates, K35. C3091: streptomycin-resistant UTI isolates, K16. MH258: MDR blood isolates, ST258.

CONCLUSION

A comparatively large genome and genetic flexibility enable K. pneumoniae to adapt to nearly any environment and harbor and disseminate ARGs (5, 21, 22, 208). Due to the threat of K. pneumoniae in a hospital setting, research has primarily focused on disease states and AMR. While new antibiotics, decontamination protocols, and quarantine procedures have been developed over the decades to combat the growing AMR and virulence of K. pneumoniae, other stages of this pathobionts life cycle have been largely neglected. As asymptomatic carriage remains the critical first step before extra-intestinal infections, further investigations using various murine models that can account for human genetic variations are necessary to elucidate the complex mechanisms and interactions of K. pneumoniae in the GI tract. Although many bacterial factors that are important during antibiotic pressure have been identified and characterized, there exists a considerable gap in our understanding of K. pneumoniae’s interactions with the gut microbiome. Thus, future studies should take these gaps in our understanding of K. pneumoniae pathogenesis into account. By focusing on this early stage of K. pneumoniae infection, we may unveil pathways to prevent or reduce silent reservoirs from transmitting this dangerous pathogen to vulnerable populations.