Background

The Raoultella genus comprises gram-negative, oxidase-negative, facultative anaerobic bacteria within the Enterobacteriaceae family. R. planticola, earlier known as Klebsiella planticola and Klebsiella trevisanii, is a gram-negative, rod-like bacterium first described by Ferragut [1] from aquatic and soil isolates and later differentiated from Klebsiella after phylogenetic analysis by Drancourt and associates [2]. Matrix-assisted laser desorption/ionization–time of flight (MALDI-TOF) mass spectrometry is the current method used to identify and differentiate R. planticola [3].

R. planticola is an emerging pathogen which has been linked to fatal infections. Only 33 cases of R. planticola were reported prior to the middle of 2015 [4]. There have been 19 novel cases reported since that time. Pediatric cases, although extremely rare, have also been reported [5]. Further, R. planticola may cause bacteremia, pneumonia, intra-abdominal infections, urinary tract infections, soft tissue infections, and conjunctivitis [6,7,8]. To our knowledge, there have been no reports of R. planticola biliary tract infections or a thorough investigation of microbial virulence or immune escape. Herein, we report a case of R. planticola biliary tract infection as a long-term postsurgical complication in a 31-year-old woman who initially presented with acute cholecystitis. We also discuss the results of a genome-based comparison between R. planticola and K. pneumoniae in order to examine possible virulence factors of R. planticola.

Case presentation

A 31-year-old Caucasian woman presented to the emergency department with sudden onset of abdominal pain, fever, chills, and malaise. She had a history of laparoscopic cholecystectomy in 2008 complicated by bile leak requiring biliary stents. In 2011, she developed hepatic cysts, which were surgically extirpated in 2011 and 2012. She remained afebrile and mostly asymptomatic, with only occasional mild right upper quadrant pain until the current presentation which caused her to seek medical attention at the emergency department. Importantly, she denied any history of solid organ transplants, hematologic malignancy, chemotherapy, transplantation neutropenia, cirrhosis, seafood ingestion, or proton pump inhibitor (PPI) use.

Vital signs on presentation were temperature of 98.1 °F, heart rate of 85 beats per minute, respiration of 18 breaths per minute, and blood pressure of 126/83 mmHg. Physical examination revealed an afebrile, anicteric female in moderate, painful distress with slight, diffuse abdominal tenderness on palpation. Laboratory/radiographic tests revealed a white blood cell count of 41.1 cells/μL; elevated liver enzymes (alanine aminotransferase 102 U/L and aspartate aminotransferase 74 U/L); alkaline phosphatase of 318 U/L; and total bilirubin of 2.4 mg/dl. Computed tomography and magnetic resonance cholangiopancreatography (MRCP) were significant for dilated right intrahepatic bile duct with evidence of a surgically absent gall bladder (Figs. 1, 2, 3). During admission in the emergency room, the patient became febrile, and blood cultures (BC) were drawn. In light of the clinical picture and imaging studies, biliary sepsis and bacteremia due to intrahepatic duct stricture were suspected. The patient was admitted to the hospital and empirically started on piperacillin/tazobactam. The BC was positive for gram-negative rods in two of two peripheral BC after 24 hours. R. planticola was reported as the isolate on the third hospital day and was resistant to ampicillin and piperacillin but susceptible to ceftriaxone (microbial resistance and susceptibilities were completed by Quest Diagnostics). Therapy was changed to ceftriaxone 2 g parenterally every 24 hours, and the patient quickly improved clinically, with normalization of liver function within 3 days (hospital day 6). She was discharged on home therapy with referral for subsequent evaluation and treatment of her intrahepatic duct strictures.

Fig. 1
figure 1

Magnetic resonance cholangiopancreatography dextral sagittal view of the abdomen

Full size image
Fig. 2
figure 2

Magnetic resonance cholangiopancreatography posteroanterior view of the abdomen

Full size image
Fig. 3
figure 3

Magnetic resonance cholangiopancreatography sinistral sagittal view of the abdomen

Full size image

Case discussion and methods

Previously reported R. planticola cases

Raoultella planticola is an emerging bacterial pathogen (see Fig. 4) that has previously been associated with nonclinical environments such as aquatic habitats and therefore has been linked to consumption of seafood [9]. However, case studies since 1985 have indicated an increased number of clinical cases [10] and multiple organ infections [11]. Previously, R. planticola was isolated from patients with comorbid leptospirosis [10] and found to be a cause of pneumonia [12]. R. planticola was also observed to occur in hematological malignancy when the organism was isolated from the oral ulcers of a patient with chemotherapy-induced oral mucositis [13]. Similarly, other cases have been documented that suggest an increased susceptibility to infection in immunocompromised states [13]. Table 1 summarizes all reported R. planticola infections prior to 2018 found during an in-depth literature search.

Fig. 4
figure 4

Timeline of reported cases associated with isolation of R. planticola

Full size image
Table 1 A summary of the epidemiology of R. planticola cases
Full size table

Analysis of all the documented patients showed that R. planticola caused bacteremia in 22% of cases, soft tissue infections in 17%, urinary tract infections in 15%, lower respiratory tract infections in 10%, and eye infections in 10%. Sources of isolation correlated with the infected organ system (r2 = 0.72). The annual timeline frequency of documented infections caused by this pathogen potentially indicate a biannual prevalence (see Fig. 4).

Pathogenesis and virulence factors in R. planticola genome

The pathogenesis of R. planticola has not yet been established; however, fimbria, biofilm production, encapsulation, lipopolysaccharide (LPS), and siderophores have been observed to be important virulence factors in the closely related species Klebsiella pneumoniae species [3, 14].

In order to investigate possible virulence factors for R. planticola, in silico analysis was conducted using the National Center for Biotechnology Information (NCBI) Basic Local Alignment Search Tool (BLAST) (NCBI.gov) [15], Universal Protein Resource (UniProt; uniprot.com) [16], and Integrated Microbial Genomes (IMG; img.jgi.gov) system [17]. In the analysis, previously documented virulence factor gene sequences, including known K. pneumoniae virulence factor gene sequences, were used to generate queries to blast against the R. planticola genome. Using these databases, gene sequences and accession numbers for virulence factors in K. pneumoniae were found for over 40 genes pertaining to pili components, pili chaperone proteins, biofilm synthesis and initiation, capsule assembly, capsule biosynthesis, capsule initiation, LPS synthesis, outer membrane surface protein chaperones, and expression of mucoid phenotype (genes present, see Tables 2, 3). Using the gene function on IMG [17], permanent drafts of K. pneumoniae pKP469IL Plasmid (B) [P], K. pneumoniae pKP531IL (B) [P], K. pneumoniae 1162281 (B) [P], and K. pneumoniae 1191100241 (B) [P] were searched for gene sequences and accession numbers for known virulence factors [18]. The genome for R. planticola strain GODA was selected as a target for querying the virulence factors relevant for endotoxins, capsules, fimbria, pili production, and biofilm production using the BLASTn and BLASTp tools available from NCBI.gov [15]. In order to investigate the clinical relevance, the genome of ATCC 33531 [19] was compared against GODA and found to be 99% identical (see Table 4).

Table 2 Genomic identification of virulence factors in the genome of R. planticola ATCC 33531
Full size table
Table 3 The results of queries previously functionally annotated in R. planticola
Full size table
Table 4 Information for the host-associated isolated ATCC 33531 strain and the environment-isolated GODA strain
Full size table

For functional gene annotation, greater than 60% query coverage, 70% nucleotide identity, and Expect (E)-value below 0.001 were used as the minimum similarity criteria between functionally documented genes of K. pneumoniae and unknown R. planticola genes (see Tables 2, 3). Novel queries identified by BLASTn analysis of R. planticola GODA against K. pneumoniae are shown in Table 2, and previously functionally annotated queries are shown in Table 3.

Results and discussion

The results of blasting for capsule and mucoid production genes are shown in Tables 2 and 3.

Endotoxin production

Wzze [20] and O-acetyltransferase [21], genes known to be involved in the synthesis of LPS, showed 93% and 82% identity in 67% and 71% of the gene fragments, respectively (see Table 2). This indicates that those mechanisms might be shared between R. planticola and K. pneumoniae and that R. planticola may produce LPS endotoxin, which could explain how it causes a bacteremia, avoids immune response, or causes sepsis.

Capsule and mucoid production

We observed the following virulence factor relationships between the target genomes: Wzi (99% identity in 94% of the gene fragments); rcsA (69% identity in 100% of the gene fragments); rcsB (82% identity of 100% of the gene fragments); KpsS (99% identity in 5% of the gene fragments) suggesting a shared prosthetic group; uncharacterized protein (99% identity in 1% of the gene fragments) suggesting a shared prosthetic group. The presence of such genetic similarities indicates that R. planticola has the potential to synthesize, regulate, and assemble a capsule and express a mucoid phenotype that could help it escape the host immune system [22,23,24,25,26]. Mucoid phenotype regulators rmpA and rmpA2 did not show significant matching, indicating that capsular mucoid composition in R. planticola may be expressed through a different mechanism from that in K. pneumoniae. The results of blasting for capsule and mucoid production genes are shown in Tables 2 and 3.

Adhesins, pili, and fimbria

Skp and SurfA queries were found to be 92% and 87% identical, respectively, between R. planticola and K. pneumoniae, suggesting that R. planticola codes for periplasmic chaperoning of outer membrane protein assembly [27, 28]. Type 1 fimbria regulatory proteins FimB and FimE and fimbria FimD queries have previously been documented in R. planticola (see Table 3). No other similarities for different fimbrial components were found. This variability that R. planticola possesses in its fimbria may contribute to its host cell attachment and opsonization prevention in ways that differ from those mechanisms in K. pneumoniae.

Biofilm production

Raoultella planticola was found to possess genes similar to K. pneumoniae that have previously been shown to cause infection via biofilm production, upregulation of motility factors, outcompeting host cells for iron, and mucous production (see Table 3).

Gene cassettes

Whole genome blasting analysis between host-associated ATCC 33531 and GODA was conducted to identity chromosomal cassettes. We observed identity of 99% in 93% of gene fragments, suggesting high gene density that could allow efficient regulation of gene expression in mechanisms of antibiotic resistance, host immune system evasion, host cell attachment and invasion, and intracellular survivability (see Table 4) [29]. R. planticola may contain pathogenicity islands, a potential result of transduction that would also be involved in bacterial adaption, but further research is needed to confirm this.

Organ systems affected, virulence factors, and potential latent infection in our case

An in-depth search of the current literature did not yield other case studies with a similar isolation of R. planticola. We were also unable to identify another case of R. planticola isolated after endoscopic retrograde cholangiopancreatography (ERCP) as part of biliary sepsis not associated with gastroenteritis. Further, a detailed history and chart review of our patient did not show any of the previously reported risk factors associated with R. planticola including bacteremia/sepsis of the gastrointestinal tract (GI), biliary malignancy, chemotherapy, transplantation, neutropenia, cirrhosis, seafood ingestion, or PPI usage. It is possible that our patient had recently become infected with R. planticola rather than during the time of her laparoscopic procedure; however, as discussed above, an extensive attempt to document any previously associated risk factors failed to illuminate any. It has been reported that patients with chronic biliary strictures are at increased risk of cholangitis, possibly due to static biliary fluid in the stenotic biliary system or because of abnormal anatomic morphology that facilitates bacterial adhesion and colonization [30,31,32].

Therefore, we speculate that our patient’s bacteremia, which developed 5 years postoperatively, may be due to possible latency of the pathogen. The in silico results might also indicate that this organism survived exposure to the host immune system through the employment of an intracellular biofilm or in a non-culturable but viable (NCBV) state for the 5-year period [33, 34]. Our results also suggest that biofilms, enterobactin, and mucoid phenotype are likely virulence factors in the pathogenesis of R. planticola’s ability to cause infection. Additionally, we identified a conservation of genes involved in pili synthesis regulation, fimbrial protein chaperoning, capsule biosynthesis, and endotoxin production; however, the genetic variation of genes coding for pili, fimbria, and capsule polysaccharide composition may indicate that these genes are subject to antigenic variation or reductive evolution in an attempt to avoid the host immune system. Multiple genomes of newly isolated clinical P. planticola should be sequenced in order to evaluate its level of evolutionary conservation of the extracellular and surface glycoproteins.

Conclusions

This unique case adds to the literature on the GI affinity of R. planticola and, with the results of the in silico analysis, suggests that potential novel risk factors for infection may be biliary tract stricture, cholecystitis, and prior surgery.