Kingella kingae, a normal component of the upper respiratory flora, is being increasingly recognized as an important invasive pathogen in young children. Genetic diversity of this species has not been studied.
We analyzed 103 strains from different countries and clinical origins by a new multilocus sequence-typing (MLST) schema. Putative virulence gene rtxA, encoding an RTX toxin, was also sequenced, and experimental virulence of representative strains was assessed in a juvenile-rat model.
Thirty-six sequence-types (ST) and nine ST-complexes (STc) were detected. The main STc 6, 14 and 23 comprised 23, 17 and 20 strains respectively, and were internationally distributed. rtxA sequencing results were mostly congruent with MLST, and showed horizontal transfer events. Of interest, all members of the distantly related ST-6 (n = 22) and ST-5 (n = 4) harboured a 33 bp duplication or triplication in their rtxA sequence, suggesting that this genetic trait arose through selective advantage. The animal model revealed significant differences in virulence among strains of the species.
Citation: Basmaci R, Yagupsky P, Ilharreborde B, Guyot K, Porat N, Chomton M, et al. (2012) Multilocus Sequence Typing and rtxA Toxin Gene Sequencing Analysis of Kingella kingae Isolates Demonstrates Genetic Diversity and International Clones. PLoS ONE 7(5): e38078. doi:10.1371/journal.pone.0038078
Editor: Christophe Sola, Institut de Génétique et Microbiologie, France
Received: February 2, 2012; Accepted: April 30, 2012; Published: May 31, 2012
Copyright: © 2012 Basmaci et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported in part by the “Académie Nationale de Médecine”. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. No additional external funding received for this study.
Competing interests: The authors have declared that no competing interests exist.
Increasing use of improved culture methods and sensitive nucleic acid amplification assays in recent years has resulted in the recognition of Kingella kingae as a common etiology of skeletal system infections and bacteremia below the age of 3 years –. The organism is carried asymptomatically in the oropharyngeal mucosal surfaces from which it may be transmitted from-person-to-person and/or penetrate into the bloodstream and disseminate to remote sites such as bones, joints or the endocardium – and, more rarely, to the meninges or the peritoneal cavity –.
Because the organism was traditionally considered an exceptional cause of human disease, our current knowledge on the population structure of the species and potential differences in terms of virulence between strains is still limited. A study by Lehours demonstrated polymorphism of the rtxA gene, which encodes the RTX toxin . Genotyping of 240 strains derived from healthy carriers by pulsed-field gel electrophoresis (PFGE) demonstrated 40 distinct clones of which a few were undistinguishable from those detected among isolates from patients with invasive infections . Although these preliminary results indicate that K. kingae exhibits genetic heterogeneity, development of a highly reproducible genotyping method is necessary to decipher the genetic organization of the species and determine whether certain genotypes are associated with either carriage or disease. Furthermore, to date the differences in pathogenicity of K. kingae strains have only been studied in vitro and no animal model of invasive infection has been developed. Our animal model may contribute to discriminate between highly and low virulent clones and would point to clones to be further investigated to identify genetic determinants of virulence.
In the present study we studied the genetic diversity of 103 K. kingae isolates from different countries and different clinical sources by use of a new multi-locus sequence-typing (MLST) scheme. We also analysed the relationship between MLST clustering and the diversity of the rtxA encoding gene. Finally we developed an animal model to assess the virulence of K. kingae isolates.
Materials and Methods
To develop an adequate MLST scheme, a diverse collection of K. kingae organisms was assembled to include strains derived from individuals living in different geographic locations and isolated in different periods, patients with a variety of clinical syndromes, as well as epidemiologically-unrelated healthy carriers.
Seven K. kingae strains, available in international reference collections (4 from France and 3 from Norway), were obtained from the Pasteur Institute of Paris (Table S1). In addition, 25 clinical isolates from 3 different French hospitals were also studied. These isolates derived from 23 patients with osteoarticular infections (OAI), bacteremia (n = 1) and endocarditis (n = 1) (Table S2).
Over the years, the Clinical Microbiology Laboratory of the Soroka University Medical Center (SUMC) of southern Israel has gathered over 1,000 K. kingae isolates from respiratory carriers living in the region, as well as 170 organisms from Israeli patients with a variety of invasive K. kingae infections. To assure a wide selection of Israeli strains, 69 candidates for MLST typing were chosen on the basis of their dissimilar PFGE profiles and clinical and epidemiological data partially published elsewhere . Whenever it was possible, pairs of invasive and colonizing isolates exhibiting indistinguishable PFGE profiles were selected. Overall, 69 Israeli isolates were studied, including 33 derived from asymptomatic carriers, 17 from patients with OAI, bacteremia (n = 10), and endocarditis (n = 9) (Table S3).
Two isolates from the USA (kindly provided by Ruth Lynfield and Joseph W. St. Geme), and one isolate from a healthy Russian pediatric carrier, were also studied. (Table S3).
Organisms were cultured on chocolate Isovitalex agar or Columbia blood agar for 24 to 48 h at 37°C in a 5% CO2-enriched athmosphere.
DNA extraction, amplification and sequencing
DNA was extracted from specimens with the BioRobot EZ1 workstation using the EZ1 DNA tissue kit (Qiagen, Courtaboeuf, France) according to the manufacturer's recommendations, and stored at −80°C.
For DNA amplification, the PCR mixture contained 25 µL of AmpliTaq Gold PCR MasterMix (Life Technologies, Villebon sur Yvette, France), 2 µL (10 µM) of each primer (forward and reverse), 19 µL of pure water and 2 µL of DNA (∼5 ng) in a final volume of 50 µL. Amplification was performed on an iCycler (Bio-Rad, Marnes la Coquette, France), with an initial step of 15 min at 95°C, followed by 35 cycles of 15 s at 95°C, 30 s at 56°C, 1 min 30 s at 72°C, and a final extension step of 10 min at 72°C. Amplification products were stored at 4°C until further tested. Amplification products were visualized under U.V. after migration on gel electrophoresis, and both strands were sequenced by a Beckman-Coulter Genomics instrument (Takeley, Essex, United Kingdom).
At the beginning of the study no Kingella species had been sequenced with the exception of one K. oralis isolate. We first selected 7 housekeeping genes (abcZ, adk, aroE, cpn60, gdh, recA and fumC) that were known for K. kingae (cpn60) , or used for related species MLST such as N. meningitidis (abcZ, adk, aroE, fumC and gdh) or H. influenzae (recA) and that were common to K. oralis. Consensus primers were designed by alignment of N. meningitidis and K. oralis sequences, available on database. Then they were tested on 4 K. kingae strains (ATCC 23330, ATCC 23331, CIP 73.1 and CIP 102473) and amplification products were sequenced, allowing us to design new consensus primers for each K. kingae gene.
In addition to the MLST scheme, the rtxA gene was also amplified using primers described by Lehours et al. and sequenced .
Estimation of the between-strains relatedness
A different allele number was given to each distinct sequence within a locus, and a distinct sequence-type (ST) number was attributed to each distinct alleles combination. Isolates were grouped into ST-complexes (STc) if they differed at no more than one locus from at least one other member of the group. Founder genotypes of STcs were defined as the ST of the STc with the highest number of neighboring STs (single locus variants). Molecular epidemiological data were stored and analysed using software BioNumerics (Version 6.6; Applied-Maths, Belgium). Relatedness between the different STs was investigated based on comparison of allelic profiles using the minimum spanning tree (MStree) method from BioNumerics. The relatedness between each rtxA sequence was shown as a dendrogram, constructed by the UPGMA method using MEGA 3.1 program .
After 24 h of culture on chocolate Isovitalex agar strains were suspended in phosphate-buffered saline (PBS) at an optic density at 600 nm of 1.2, corresponding to a concentration of approximately 108 CFU/mL. Determination of bacterial concentration was retrospectively performed by serial dilution and plating.
One hundred µL of these suspensions (107 CFU) were injected intraperitoneally in the left inferior quadrant of 5-day-juvenile albino Spragues Dawley rats (Charles River Laboratories, France). Inoculated animals were examined daily for 5 consecutive days for skin discoloration, signs of peritonitis and survival rates. If subcutaneous lesions were observed, a piece of tissue was removed, minced, suspended in 500 µL of brain-heart infusion and plated onto chocolate agar.
All experiments and animals housing procedures complied with the University Paris Diderot animal ethics committee guidelines and were approved by local animal ethics committee “Comité d'Éthique Expérimentation Animale Bichat-Debré” (2011-14/676-0054).
Housekeeping genes characteristics (Table 1)
Among the 7 genes initially selected, fumC was abandoned because we did not succeed to amplify it in K. kingae. Therefore, 6 couples of primers were designed after analyzing the sequence of the target genes (abcZ, adk, aroE, cpn60, gdh and recA) and allowed amplification of all but 6 strains. For 5 of them, an alternative cpn60-forward-primer was designed and allowed sequencing. Only one strain (PV1572 from a healthy carrier) was unsuccessfully sequence-typed by failing of aroE amplification. Lengths of analyzed sequences were between 303 bp (cpn60) and 576 bp (gdh). The number of alleles is indicated in Table 1 and ranged from 6 (recA) to 16 (abcZ). Allele sequences of each gene for all strains are available on the Pasteur Institute of Paris website http://www.pasteur.fr/recherche/genopole/PF8/mlst/Kingella_kingae.html.
Estimating relatedness between the strains
The analysis of the sequences of the 6 housekeeping genes for the 103 K. kingae isolates revealed 36 STs, among them 21 were unique (Table 2). The 3 most common STcs (STc-6, STc-14 and STc-23) included 60/103 (58.3%) of all typed isolates. These three STcs were distantly related (Figure 1), as STc-14 shared no common allele with the two others and STc-6 shared only one common allele with STc-23. STc-6 was mainly composed of one ST (ST6, n = 22; ST7, n = 1) whereas STc-23 (n = 20) and STc-14 (n = 17) were more diverse and included STs-21-22-23-24, and STs-14-15-16-17-18, respectively. Of note, the type strain (ATCC 23330) was the only strain of ST-1 which belong to STc-1 (n = 4) and was distantly related to the most frequent STcs (Figure 1).
Each circle corresponds to a sequence type (ST). The ST number is given beside the circle, and the size of the circle is related to the number of isolates found with that profile (from 1 for small circles e.g. ST-1 to 22 for ST-6). Each colour inside circles represents geographical origin of the strains (green: Israel; red: France; blue: Norway; yellow: USA and turquoise: Russia). Colored zones between some groups of circles indicate that these profiles belong to the same ST complex. The three major ST complex are indicated. Width of line joining two STs indicates the number of alleles differing.
Some STcs were worldwide distributed. STc-6 included 18 isolates from Israel and 5 from France, and STc-14 and STc-23 included isolates from Israel, France, Norway, and the USA, whilst some STcs seemed to be country-specific, such as the exclusive French STc-25, representing 7/29 (24.1%) of the French strains (Table 2 and Figure 1).
Polymorphism of K. kingae rtxA gene
To examine the potential relationships between sequence polymorphism of the core genome and that of a putative virulence gene, we studied sequences of the haemolysin gene rtxA and compared it with the MLST results. A sequence ranging from 909 to 975 bp was analyzed for the gene rtxA for the 103 K. kingae strains.. The mean G+C% of this gene was strikingly different from that of the MLST schema (37.6% vs. 49.7% respectively) and from that of the largest contig (47,149 bp) of the recently sequenced K. kingae type strain (accession number AFHS01000014.1, G+C% = 48.1%). The genetic relatedness of rtxA sequences is shown as a dendrogram (Figure 2) and in Table 2. This gene was remarkably polymorphic and exhibited 18 different sequences. The 18 sequences are available on NCBI database under accession number JQ340459 to JQ340476. Four rtxA alleles (15 to 18) found in 10 strains, constituted a subgroup distantly related to the 14 others (Figure 2). This may suggest that the 10 strains had acquired their rtxA genes from a source different from that of other K. kingae strains. However none of the rtxA alleles had a significant homology to sequences in GenBank database. rtxA sequencing results were congruent in most strains with their allocation to ST and STc by MLST. For instance, rtxA-allelles 8 to 11 included all and exclusively strains of the major ST-6 (Figure 2). However, incongruence was observed for several alleles: the rtxA-allele 14 was dispersed in unrelated ST-3 and STc-14, and the rtxA-allele 1 was distributed among the distantly related ST-2, ST-8, STc-23 and STc-25; the strain belonging to ST-17 was the only member of STc-14 that had the same rtxA allele found in STc-29. Finally, the 4 rtxA alleles (15 to 18) constituting a subgroup, were scattered among unrelated STs. Altogether, these results suggest that horizontal transfers of rtxA alleles had occurred in K. kingae species however the random occurrence of multiple advantageous mutations in the rtxA genes in strains belonging to distantly related STs can not be excluded.
Distribution of 103 strains according to their ST among the 18 alleles is also indicated. All the strains with alleles 8, 10, 11 and 17 have a duplication of 33 bp, and all allele 9 strains have a triplication (see Figure 3).
Of note, in 26 rtxA sequences, we observed an insertion of 33 bp, corresponding to 11 amino-acids, which was repeated twice in two isolates (HAM and KK6) (Figure 2 and 3). Interestingly, all these strains belonged to two unrelated STs, ST-6 and ST-5 (Figure 1). The location and the sequence of this insertion were identical for all strains, except HAM and KK6, and appeared in nucleotide 76 (Figure 3). This insertion is a quasi perfect duplication of the immediately previous 33 bp, corresponding to following aminoacid sequence QAG(V/A)QALN(R/K)AG (Figure 3). Moreover, rtxA sequences were very different between strains of ST-5 and ST-6 (Figure 2). These observations are not in favour of a horizontal transfer of rtxA genes between these two STs, which could have explain their similar particular genetic traits, but plead for the occurrence of two unrelated insertion events.
For each strain, first line represents nucleotide-sequence; second line represents aminoacid sequence. Lines 1 and 2: type strain ATCC 23330; lines 3 and 4: strain BB631 of ST-5 with 33 bp duplication; lines 5 and 6: strain SIL of ST-6 with 33 bp duplication; lines 7 and 8: HAM of ST-6 with 33 bp triplication. Mismatches compared to type strain are indicated in red; the inserted sequences are indicated in blue and underlined; the original sequence, which is duplicated, is in purple. Mismatches between duplicated-inserted sequences and the corresponding original sequences are indicated in green.
To determine whether differences of virulence exist between K. kingae isolates, we infected three groups of eight 5-day-juvenile-rats with 3 MLST-unrelated-isolates of different clinical origin. We selected the strains ATCC 23330 (type strain, isolated from the nose, unique strain of the ST-1), HAN (isolated from a patient with OAI and belonging to the major ST-6) and N10770 (from a patient with bacteremia, belonging to ST-28, and representing a minor invasive ST). The inocula injected intraperitoneally were 107 CFU for each strain. The experiment was repeated twice and results are presented pooled. The disease-free survival curve showed significant differences (Figure 4). Animals infected with strain ATCC 23330 did not present any sign of infection at any time and were all alive and healthy at day 5. In contrast, in the N10770 inoculted rats, 10 animals were dead at 24 h, and five of the six surviving animals had a major ischemic-necrosis lesion in the abdomen that engulfed their left hind leg. These five animals were then sacrificed. The remaining rat was alive and healthy at day 5. The HAN strain displayed an intermediate virulence, all the 16 animals were alive at 48 h, but 6 had clinical symptoms of infection: five had an abscess on the abdomen and one looked ill. Aspiration of the abscesses revealed polymorphonuclear leucocytes and viable K. kingae organisms. Four animals died between days 3 and 5, and 8 animals remained healthy.
To date, the population structure of the emerging pathogen K. kingae remains largely unknown. In this study we established, for the first time, a MLST schema of the species and observed a wide genetic diversity with 36 distinct STs and 9 STcs among 103 isolates. We demonstrated that 3 distinct STcs (STc-6, STc-14 and STc-23), sharing almost no common alleles and including 58.3% of all isolates, were represented in geographically distant countries, suggesting that these strains probably posses a selective biologic advantage and have evolved independently.
Whilst the difference between the G+C% of core genome and rtxA suggest that the haemolysin-encoding gene had been acquired from a different bacterial species, important congruencies observed suggest a remote acquisition and co-evolution of the rtxA gene and the genes examined for MLST typing. We also noted a duplication of 33 bp in some sequence types of the rtxA gene, which were previously described by Lehours et al. . Interestingly, duplications occurred exclusively among ST-6 and ST-5 strains. However, we made a different interpretation of the genetic event to that Lehours et al. Indeed, we found that this insertion occurred in position 76 and not in position 67, and, although a single non-synonymous mutation was detected, we concluded that this insertion is a duplication of the immediately adjacent 33 bp fragment. Furthermore, considering that sequences out of the insertion site were different between both STs, we hypothesized that this genetic trait arose through convergent evolution, and probably not by horizontal transfer of the rtxA gene, confering a selective biological advantage. Whether the genetic structure of this 33 bp sequence promotes duplication independently of any selective advantage remains to be determined.
To our knowledge, we developed the first experimental model of K. kingae infections. This model of intraperitoneal infection in juvenile rats demonstrated clear differences in virulence between wild strains and the ATCC 23330 type strain, which exhibited the lowest virulence. Whereas this organism isolated in the 1960's from the respiratory tract was originally less invasive or rather experienced genome reduction because of repeat subcultures over the years is unknown. The two other strains examined exhibited intermediate and high degree of virulence. Therefore, the herein described animal model can serve to identify virulent clones among candidates from epidemiological data in order to determine novel potential virulence factors.
In conclusion, our MLST tool will serve future epidemiological studies and transcontinental comparisons providing a comprehensive view of circulating clones and detecting emerging ones. Moreover, the MLST data may be used for targeting clones to be further investigated for virulence in our animal model. We also suggest that the recent fully sequenced ATCC 23330 type strain, which appears devoid of pathogenicity in our animal model, may serve for future comparative genomic studies.
Strains of Kingella kingae available on international collections used in the study.
Twenty-five French clinical Kingella kingae isolates.
Seventy-one Kingella kingae isolates, typed by MLST from Israel, Russia and USA.
We thank Dr. Jocelyne Caillon, Dr. Hoang Vu-Thien and Dr. Eric Vallée for providing their K. kingae isolates from France. We thank Ruth Lynfield and Joseph W. St Geme for providing their K. kingae isolates from USA. We thank Valerie Caro and Sylvain Brisse from the Pasteur Institute of Paris for creating and hosting the public database.
Conceived and designed the experiments: PB SB EB. Performed the experiments: RB KG MC PB. Analyzed the data: RB JMT SB EB PB. Contributed reagents/materials/analysis tools: PY NP BI KM. Wrote the paper: RB EB SB PB.
- 1. Yagupsky P, Dagan R, Howard CW, Einhorn M, Kassis I, et al. (1992) High prevalence of Kingella kingae in joint fluid from children with septic arthritis revealed by the BACTEC blood culture system. J Clin Microbiol 30: 1278–1281.
- 2. Yagupsky P (2004) Kingella kingae: from medical rarity to an emerging paediatric pathogen. Lancet Infect Dis 4: 358–367.
- 3. Cherkaoui A, Ceroni D, Ferey S, Emonet S, Schrenzel J (2009) [Pediatric osteo-articular infections with negative culture results: what about Kingella kingae?]. Rev Med Suisse 5: 2235–2239.
- 4. Verdier I, Gayet-Ageron A, Ploton C, Taylor P, Benito Y, et al. (2005) Contribution of a broad range polymerase chain reaction to the diagnosis of osteoarticular infections caused by Kingella kingae: description of twenty-four recent pediatric diagnoses. Pediatr Infect Dis J 24: 692–696.
- 5. Stahelin J, Goldenberger D, Gnehm HE, Altwegg M (1998) Polymerase chain reaction diagnosis of Kingella kingae arthritis in a young child. Clin Infect Dis 27: 1328–1329.
- 6. Rosey AL, Abachin E, Quesnes G, Cadilhac C, Pejin Z, et al. (2007) Development of a broad-range 16 S rDNA real-time PCR for the diagnosis of septic arthritis in children. J Microbiol Methods 68: 88–93.
- 7. Ilharreborde B, Bidet P, Lorrot M, Even J, Mariani-Kurkdjian P, et al. (2009) New Real-Time PCR-Based Method for Kingella kingae DNA Detection: Application to Samples Collected from 89 Children with Acute Arthritis. J Clin Microbiol 47: 1837–1841.
- 8. Chometon S, Benito Y, Chaker M, Boisset S, Ploton C, et al. (2007) Specific real-time polymerase chain reaction places Kingella kingae as the most common cause of osteoarticular infections in young children. Pediatr Infect Dis J 26: 377–381.
- 9. Ceroni D, Cherkaoui A, Ferey S, Kaelin A, Schrenzel J (2010) Kingella kingae osteoarticular infections in young children: clinical features and contribution of a new specific real-time PCR assay to the diagnosis. J Pediatr Orthop 30: 301–304.
- 10. Lehours P, Freydiere AM, Richer O, Burucoa C, Boisset S, et al. (2011) The rtxA Toxin Gene of Kingella kingae: a Pertinent Target for Molecular Diagnosis of Osteoarticular Infections. J Clin Microbiol 49: 1245–1250.
- 11. Basmaci R, Lorrot M, Bidet P, Doit C, Vitoux C, et al. (2011) Comparison of clinical and biologic features of Kingella kingae and Staphylococcus aureus arthritis at initial evaluation. Pediatr Infect Dis J 30: 902–904.
- 12. Yagupsky P, Porsch E, St Geme JW 3rd (2011) Kingella kingae: An Emerging Pathogen in Young Children. Pediatrics 127: 557–565.
- 13. Dubnov-Raz G, Ephros M, Garty BZ, Schlesinger Y, Maayan-Metzger A, et al. (2010) Invasive pediatric Kingella kingae Infections: a nationwide collaborative study. Pediatr Infect Dis J 29: 639–643.
- 14. Sena AC, Seed P, Nicholson B, Joyce M, Cunningham CK (2010) Kingella kingae endocarditis and a cluster investigation among daycare attendees. Pediatr Infect Dis J 29: 86–88.
- 15. Youssef D, Henaine R, Di Filippo S (2010) Subtle bacterial endocarditis due to Kingella kingae in an infant: a case report. Cardiol Young 20: 448–450.
- 16. Holmes AA, Hung T, Human DG, Campbell AI (2011) Kingella kingae endocarditis: A rare case of mitral valve perforation. Ann Pediatr Cardiol 4: 210–212.
- 17. Toshniwal R, Draghi TC, Kocka FE, Kallick CA (1986) Manifestations of Kingella kingae infections in adults: resemblance to neisserial infections. Diagn Microbiol Infect Dis 5: 81–85.
- 18. Hay F, Chellun P, Romaru A, d'Auzac P, Vidal MP, et al. (2002) [Kingella kingae: a rare cause of meningitis]. Arch Pediatr 9: 37–40.
- 19. Bofinger JJ, Fekete T, Samuel R (2007) Bacterial peritonitis caused by Kingella kingae. J Clin Microbiol 45: 3118–3120.
- 20. Yagupsky P, Weiss-Salz I, Fluss R, Freedman L, Peled N, et al. (2009) Dissemination of Kingella kingae in the community and long-term persistence of invasive clones. Pediatr Infect Dis J 28: 707–710.
- 21. Kumar S, Tamura K, Nei M (2004) MEGA3: Integrated software for Molecular Evolutionary Genetics Analysis and sequence alignment. Brief Bioinform 5: 150–163.