http://orcid.org/0000-0001-6458-3530
Figures
Abstract
Background
The Gram-negative bacterium Kerstersia gyiorum, a potential etiological agent of clinical infections, was isolated from several human patients presenting clinical symptoms. Its significance as a possible pathogen has been previously overlooked as no disease has thus far been definitively associated with this bacterium. To better understand how the organism contributes to the infectious disease, we determined the complete genomic sequence of K. gyiorum SWMUKG01, the first clinical isolate from southwest China.
Results
The genomic data obtained displayed a single circular chromosome of 3, 945, 801 base pairs in length, which contains 3, 441 protein-coding genes, 55 tRNA genes and 9 rRNA genes. Analysis on the full spectrum of protein coding genes for cellular structures, two-component regulatory systems and iron uptake pathways that may be important for the success of the bacterial survival, colonization and establishment in the host conferred new insights into the virulence characteristics of K. gyiorum. Phylogenomic comparisons with Alcaligenaceae species indicated that K. gyiorum SWMUKG01 had a close evolutionary relationships with Alcaligenes aquatilis and Alcaligenes faecalis.
Citation: Li Y, Tang M, Wang G, Li C, Chen W, Luo Y, et al. (2019) Genomic characterization of Kerstersia gyiorum SWMUKG01, an isolate from a patient with respiratory infection in China. PLoS ONE 14(4): e0214686. https://doi.org/10.1371/journal.pone.0214686
Editor: Yung-Fu Chang, Cornell University, UNITED STATES
Received: February 13, 2019; Accepted: March 18, 2019; Published: April 12, 2019
Copyright: © 2019 Li 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.
Data Availability: All relevant data are within the paper and its Supporting Information files. The genome sequence data of K. gyiorum SWMUKG01 were deposited in GenBank with the accession number CP033936. Raw sequencing data were deposited under a BioProject with accession number PRJNA497911.
Funding: This work was supported by Project of Education Department in Sichuan, China (18ZB0633), Natural Science Foundation of Southwest Medical University (No.2017-ZRZD-022 and 2018-ZRZD-011), and National Undergraduate Innovation and Entrepreneurship Project (No.201816032021). The funders had no role in the design of the study and collection, analysis, and interpretation of data and in writing the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Kerstersia gyiorum is a Gram-negative coccobacillus that is occasionally isolated from clinical samples of human infections. On nutrient agar, colonies of K. gyiorum are usually characterized by spreading edge morphology, exhibiting flat or slightly convex with smooth margins [1–3]. The word ‘gyiorum’, meaning ‘from the limbs’, was given as a species name by Coenye et al., since the organism was primarily isolated from lower-extremity wounds. The novel genus Kerstersia, initially described by Coenye et al., is grouped in the family Alcaligenaceae along with Alcaligenes, Achromobacter, Bordetella, and Pigmentiphaga spp. [1]. Kerstersia members resemble Alcaligenes faecalis phenotypically, except that isolates of Kerstersia are oxidase-negative and do not produce a fruity odor.
The initial publication describing K. gyiorum by Coenye et al. in 2003 reported six isolates recovered from leg wounds, sputum, and feces. After 2012, case reports documenting Kerstersia spp. infection began to emerge again. Since then, there have been one case of infection with a second species of Kerstersia, Kerstersia similis [4] and 12 publications describing 14 cases of patients with various diseases infected with K. gyiorum [2, 3, 5–13]. Among them, seven cases were associated with chronic otitis media [2, 5, 6, 8, 11, 13], three with chronic leg wound [2, 7, 10]. Additionally, K. gyiorum was also reported to be isolated from patients with chronic tracheostomy [3], chronic osteomyelitis [12], or urinary tract infection [9]. In fact, it is difficult to distinguish K. gyiorum from other microorganisms using conventional methods, such as traditional biochemical tests and automated identification systems, which may lead to K. gyiorum being identified incorrectly or unsuccessfully in the past in most clinical laboratories [12]. The potentially clinical importance of K. gyiorum may therefore be overlooked. For the current, with the development of (matrix-assisted laser desorption/ionization time-of-flight mass spectrometry) MALDI–TOF MS and 16S rRNA gene sequencing, K. gyiorum is expected to be identified in laboratories more accurately.
As previous reports indicated, other bacterial species were co-isolated with K. gyiorum from specimens in most cases [2, 3, 6, 9]. It was suggested that K. gyiorum has an affinity towards causing chronic mixed infections in patients [6]. However, little is known regarding how much of the disease process can specifically be attributed to K. gyiorum because of a lack of pathogenesis-related information on this organism in the literature. Reported cases of human Kerstersia infections showed that the patients gained remission when treated with an antimicrobial agent to which this organism was susceptible [2, 8, 9]. These observations indicate that K. gyiorum could contribute significantly as a possible aetiology. To better understand how the organism contributes to infections, virulence factors and pathogenic mechanisms of K. gyiorum need to be investigated. Whole genome sequencing represents a valuable approach in an in-depth exploration of potential virulence factors and may answer important questions concerning the evolution of this bacterium [14]. Though a draft genome of K. gyiorum was announced by Greninger et al [15], a comprehensive genome analysis of this organism is still not available to date. For this study, we presented a complete genome sequence of recently identified isolate K. gyiorum SWMUKG01, which was recovered from the sputum of a patient with respiratory infection. We reported and analyzed the complete genome sequence of K. gyiorum SWMUKG01, accompanied by a detailed annotation of its genome organization and expression strategy, with an emphasis on the investigation of genes and operons related to potential virulence factors.
Material and methods
Bacterial strain and growth conditions
The K. gyiorum strain SWMUKG01 was recovered from the sputum of a 70-year-old female patient of respiratory infection with a history of tracheotomy and epilepsy in May, 2018 from the affiliated hospital of Southwest Medical University. The isolate SWMUKG01 was identified as K. gyiorum by MALDI-TOF MS and 16S rRNA gene sequencing. For genomic DNA extraction, one colony of K. gyiorum SWMUKG01 was transferred into 5 ml Tryptic Soy Broth (TSB, Difco Laboratories) and cultured at 37°C. After an overnight incubation, the culture was diluted 1:100 into fresh TSB for sub-cultivation until mid-exponential growth phase was reached. The bacterial cells were collected by centrifugation at 5000 g for 10 min.
DNA extraction and genome sequencing
Total genomic DNA of K. gyiorum SWMUKG01 was extracted using Rapid Bacterial Genomic DNA Isolation Kit (Sangon Biotech, Shanghai, China) according to the manufacturer’s protocol. The genomic DNA was sent to Sangon Biotech (Shanghai, China) for de novo whole genome sequencing. A combination of HiSeq 2500 Sequencer (Illumina, San Diego, CA, USA) and the PacBio RSII platforms (Pacific Biosciences, Menlo Park, CA, USA) was employed for whole genome sequencing. De novo genome assembly of filtered reads was conducted using the Hierarchical Genome Assembly Process workflow (HGAP, v3; Pacific Biosciences) [16].
Sequence analysis
Protein coding sequences (CDSs), tRNAs and rRNAs were predicted in K. gyiorum SWMUKG01 complete genome with Prokka [17]. BLAST [18] was employed for annotation based on the sequence similarity of CDS against the Cluster of Orthologous Groups of proteins (COG) [19, 20], Kyoto Encyclopedia of Genes and Genomes (KEGG) [21], SwissProt [22], NR, and Gene Ontology (GO) databases [23]. The genome circular map was performed with the program CGview [24]. BLAST was employed to align the gene sequences against the Comprehensive Antibiotic Resistance Database (CARD) [25], and the description of the best hit (with the highest alignment length percentage and match identity) was assigned as the annotation of predicted gene. Virulence factors were investigated by BLAST against the virulence factor database (VFDB) (E < 1e-5) [26]. Genomic islands were annotated using IslandViewer 4 (http://www.pathogenomics.sfu.ca/islandviewer/) [27].
Phylogenetic analysis
For phylogenetic and comparative analysis, the genome sequences of the closest genetic relatives of K. gyiorum SWMUKG01 were obtained from the NCBI including K. gyiorum CG1 (NZ_LBNE00000000.1), A. faecalis ZD02 (NZ_CP013119.1), Alcaligenes aquatilis BU33N (NZ_CP022390.1), Achromobacter xylosoxidans NCTC10807 (NZ_LN831029.1), Achromobacter denitrificans USDA-ARS-USMARC-56712 (NZ_CP013923.1), Achromobacter insolitus FDAARGOS_88 (NZ_CP026973.1), Bordetella bronchiseptica 253 (NC_019382.1), Bordetella pertussis Tohama I (NC_002929.2), Bordetella holmesii ATCC 51541 (NZ_CP007494.1), Bordetella hinzii ASM107827v1 (NZ_CP012076.1), Bordetella petrii ASM6720v1 (NC_010170.1), Bordetella parapertussis Bpp5 (NC_018828.1), Bordetella avium 197N (NC_010645.1), Bordetella trematum H044680328 (NZ_LT546645.1), and Bordetella pseudohinzii HI4681 (NZ_CP016440.1). A phylogenetic tree based on the 16S rRNA gene sequences was generated using the tool FastTree [28]. To identify unique and conserved genes, the genome sequences of K. gyiorum SWMUKG01 and selected strains were compared using Pan-Genomes Analysis Pipeline (PGAP) [29], with which the genes shared by all genomes were collected, concatenated and aligned. The alignment of the conserved genes was used for the construction of neighbor joining tree using PGAP [29]. The species identification was also performed by average nucleotide identity (ANI) analysis between the isolate and strains shown in Table 1 using JSpeciesWS (http://jspecies.ribohost.com/jspeciesws/#analyse).
Results and discussions
General features of the genome
The genome of Kerstersia gyiorum SWMUKG01 is composed of 3, 945, 801 base pairs (bps) with a single circular chromosome (Fig 1), which showed 95% coverage and 99% identity with K. gyiorum CG1. However, in CG1, the genome is comprised of one plasmid in addition to a circular chromosome. The putative replication origin (oriC) of SWMUKG01 chromosome was identified to be located from 3, 914, 889 to 3, 915, 821 bp, with the web-based system Ori-Finder [30]. The chromosome encodes 3520 predicted genes with an average length of 994 bps, which account for 88.72% of the whole chromosome in sum. The entire SWMUKG01 chromosome contains 55 tRNA genes and 9 rRNA genes. Global characterizations of the SWMUKG01 genome are compared to those of strains CG1 (K. gyiorum), ZD02 (Alcaligenes faecalis), BU33N (A. aquatilis), NCTC10807 (A. xylosoxidans), FDAARGOS_88 (A. insolitus), 253 (B. bronchiseptica), Tohama I (B. pertussis), and Bpp5 (B. parapertussis) (Table 1).
Circles from outside to inside: CDS on the forward strand (1) and CDS on the reverse strand (2), belonging to different COG categories were indicated by different colors; rRNA and tRNA genes (3); GC content (4); GC skew with + value (green) and–value (purple) (5).
Functional classification of K. gyiorum SWMUKG01
Orthologs are thought to retain the same function during evolution. Therefore, the identification of orthologs contributes to the prediction of gene functions in a newly identified species. In this study, NCBI COG database was employed for genome-scale analysis of protein function prediction in K. gyiorum SWMUKG01. Of the 3441 protein-coding genes in SWMUKG01, 2801 were categorized into 23 COG functional codes (S1 Fig, S1 Table), but 640 were not assigned. The majority of protein-coding genes were involved in basic cellular functions, such as metabolism, transcription and translation, 34.4% of genes have unknown function, including “general function prediction only,” “Function unknown,” and the genes not assigned (S2 Table).
Phylogenetic analysis
A comparative analysis of the 16 genomes within the family Alcaligenaceae was performed to confirm the evolutionary relationship based on the 16S rRNA genes and genome-wide comparisons of orthologous gene pairs. The phylogenetic tree based on the 16S rRNA gene sequences showed that K. gyiorum SWMUKG01 and CG1 are located on the same node (98.3% similarity) (Fig 2A), both of which are more closely related to A. aquatilis and Alcaligenes faecalis, with 93.53% and 92.9% similarities of K. gyiorum SWMUKG01 in relative to A. aquatilis BU33N and A. faecalis ZD02. Genome-wide comparisons showed that 880 core genes were shared among K. gyiorum SWMUKG01 and the other 15 closely related species (Fig 3), based on which, the phylogenetic tree was reconstructed and also demonstrated that K. gyiorum SWMUKG01 and CG1 display closer evolutionary relationship to A. aquatilis and A. faecalis than to other species tested (Fig 2B). The strain SWMUKG01 was further confirmed to belong to K. gyiorum based on the ANI analysis, as it had a 99.1% ANI value with K. gyiorum CG1, which is obviously above the 95%-96% cut-off usually used to define a bacterial species [31].
(A) The phylogenetic tree was constructed using FastTree based on the 16S rRNA gene sequences. (B) The phylogenetic tree was constructed using PGAP based on 880 randomly concatenated, core genes in the genomes of K. gyiorum SWMUKG01 and 15 other closely related bacteria.
The number in the center of Venn diagram indicates the number of genes shared by all bacteria.
Analysis of virulence factors
326 potential virulence factors were annotated in strain SWMUKG01 genome and these proteins fell into 134 VF terms. Among all these potential virulence factors, proteins involved in flagella and pili production, biosynthesis of lipopolysaccharide and capsule, iron acquisition, secretion and efflux pump systems as well as two-component systems were included (S3 Table). Further studies, e.g. gene knockout studies and animal experiments, were necessary for elucidating the contribution of these virulence factors to pathogenicity of K. gyiorum. The genomic characteristics of the specific pathogenesis/virulence factors are described in detail below:
Flagella biosynthesis.
In the family of Alcaligenaceae, flagella have been widely identified in the genera of Alcaligenes, Achromobacter and Pigmentiphaga [32, 33]. In Bordetella, the flagellar operons of both B. pertussis and B. parapertussis are inactivated, leading to inability to make flagella [34]. Of the two species in Kerstersia, K. similis was reported to have no motility due to lacking flagella [4], whereas a full flagellar regulon (a cluster of operons) was identified in the genome of K. gyiorum SWMUKG01 in this study (Fig 4).
The box arrows represent ORFs.
BLASTn searches revealed that the genes of the flagellar regulon in strain SWMUKG01 showed 99% identity with those in K. gyiorum CG1, which indicates that the flagellar system in K. gyiorum is highly conserved. The putative flagellar regulon (PROKKA_00866–00919) in SWMUKG01, approximately 52 kb, encodes proteins involved in flagella biosynthesis, export, motor and bacterial chemotaxis (S4 Table). The bacterial chemotaxis system encoded by PROKKA_00877–00884 and PROKKA_00903–00904 shares 71% identity and 72% sequence coverage with those in Bordetella genomosp. 8. It is suggested that bacterial chemotaxis operates as important part of a complex network of signaling pathways, by which bacteria adjust and produce an optimal physiological response to an ever-changing environment[35]. Here, we propose that the flagellar system in K. gyiorum may contribute to its survival in a hostile environment during infection and to the generation of pathogenic responses.
Adherence.
As previously reported, tad (tight adherence) genes encode the machinery that is essential for the assembly of adhesive Flp (fimbrial low-molecular-weight protein) pili, which are required for autoaggregation, colonization, biofilm formation and pathogenesis in the genera Actinobacillus, Haemophilus, Pseudomonas, Yersinia and perhaps others [36]. Analysis of the genome sequence facilitates the identification of a tad gene cluster tadZABCD (PROKKA_01524–01528) in the strain SWMUKG01 (Fig 5), which is highly conserved in K. gyiorum showing 96% identity with that in CG1. Similar tad loci have also been identified in several Gram-negative pathogens, such as Pasteurella multocida, Yersinia pestis and Vibrio cholerae [36, 37]. Based on sequence comparisons in the family of Alcaligenaceae, the tadABC genes in K. gyiorum SWMUKG01 are highly homologous to those in some other members. For instance, the tadABC in strain SWMUKG01 showed 78%, 74% and 75% identities comparing to those of A. xylosoxidans and 78%, 74% and 70% identities with that of B. pertussis, respectively. Despite low homologies with those from other species, the putative protein encoded by tadZ in strain SWMUKG01 was predicted to be involved in localization of pilus biogenesis and TadD potentially contained a tetratricopeptide repeat protein–protein interaction motif, which is required for the assembly of Flp pili [36].
The box arrows represent ORFs.
Furthermore, we found that the flp-tadVEF-rcpCA gene region was closely linked to the tad cluster in SWMUKG01 genome, forming a 11-gene cluster flp-tadVEF-rcpCA-tadZABCD (PROKKA_01518–01528) that is predicted to be involved in the biosynthesis and secretion of Flp pili [36] (Table 2). However, flp and tadV of the cluster were missing in CG1. By a further search in the SWMUKG01 genome, we found another incomplete subset of tad operon, tadE-orf-rcpCA-tadABC (PROKKA_01139–01145) (Fig 5). Due to the lack of flp gene, which encodes the major structural component of Flp pili, this truncated tad cluster is likely to be null. It is known that the Flp pili are a distinct clade of type IVb pili [38]. To explore the existence of other types of pili in strain SWMUKG01, we searched all the genes encoding proteins for pilus biosynthesis in the genome in addition to the two tad clusters described above. As a result, a few genes, scattered throughout the SWMUKG01 chromosome, were found coding for proteins putatively involved in the type IVa pilus biosynthesis, including pilZ (PROKKA_00005), pilF (PROKKA_00008), pilD (PROKKA_00160) and pilP (PROKKA_02422). However, these CDs seem to be incomplete for the type IVa pili production, as the gene coding for the main pilus subunit pilin (PilA), a central component, was not found [38]. Besides, genes encoding potential functional analogues of type 1 or P pili are also absent in the SWMUKG01 genome.
Biosynthesis of lipopolysaccharides (LPSs).
Surface polysaccharides are extremely diverse and occur in multiple forms in Gram-negative bacteria, such as lipopolysaccharides (LPSs), the essential components of the outer membrane structurally and functionally, and capsular polysaccharides, surface layers or capsules that are associated with the cell [39].
All genes encoding enzymes for biosynthesis of the lipid A and core OS in the SWMUKG01 genome were identified (Table 3), which are highly conserved in K. gyiorum showing no less than 94% identities with those in CG1. These CDSs are scattered throughout the SWMUKG01 chromosome, just like most Gram-negative bacteria, such as E. coli, Neisseria meningitidis, Y. pestis and Pseudomonas aeruginosa. The majority of genes essential for the synthesis of lipid A (lpxK, lpxD, lpxA, lpxB, lpxC, lpxL, kdsA, kdsC and kdsD) and KDO core oligosaccharide (waaP, waaC, waaG, waaA, waaF, waaE, lpsB and gmhA) are highly conserved among different species within the family Alcaligenaceae. However, kdsB in K. gyiorum SWMUKG01, encoding CMP-2-keto-3-deoxyoctulosonic acid synthetase involved in lipid A synthesis, showed little similarity with that in other Alcaligenaceae members, but shared 71% sequence identity in Pseudomonas monteilii and Pseudomonas citronellolis. lpsE, encoding glycosyltransferase in core OS biosynthesis, showed 67% identity (98% sequence coverage) with that in Stenotrophomonas rhizophila but only 23% coverage in B. pertussis. The gene msbA (PROKKA_01899) is also identified encoding a putative lipid A export ATP-binding/permease protein (583aa) that is required for the fiipping of lipid A / core moiety of LPS from the cytoplasmic side of the IM to the periplasmic face.
9 CDSs, showing more than 96% identities with those in CG1, that are likely associated with O-antigen biogenesis were predicted in the SWMUKG01 genome. These CDSs could be classified into three groups: nucleotide sugar synthetases by a potential rfb operon (rfbCADB, PROKKA_01350–01353), glycosyltransferase genes wcaG, wecC, and wecA (PROKKA_02964–02966) and oligosaccharide repeat unit processing genes wzy (PROKKA_03380) and wzx (PROKKA_02023). Of these CDSs, only genes encoding the dTDP-glucose 4, 6-dehydratase RfbB and UDP-N-acetyl-D-glucosamine 6-dehydrogenase WecC are conserved across a wide range of species in Alcaligenaceae family, such as B. bronchiseptica and B. pertussis. It is known that in the Wzy-dependent pathway implicated in the assembly or export of O-antigen, O-unit polymerase Wzy, O-unit flippase Wzx and the O-chain length determinant Wzz are included [40]. However, we could not detect the wzz homolog in K. gyiorum SWMUKG01. Further experimental works are required to verify the O-antigen-processing process.
Capsular polysaccharides (CPS).
A cluster of genes coding for proteins involved in CPS biosynthesis and export were identified in the strain SWMUKG01 ranging from PROKKA_01910 to PROKKA_01935 (Table 4). However, several genes from PROKKA_01920 to PROKKA_01928 in this cluster were not detected in CG1, which might be due to the incomplete genome sequence. The G+C content of this gene cluster is much lower (53%) than that of the SWMUKG01 chromosome (62%), which is similar to the O-antigen’s case in many Gram-negative bacteria [41]. In fact, the genetic loci for CPS production in K. gyiorum SWMUKG01 seem to be allelic to many LPS biosynthetic loci, and multiple enzymes are the same between these pathways, such as flippase Wzx, glycosyltransferase RfaG and UDP-N-acetyl-D-mannosaminuronate dehydrogenase WecC. It is known that Wza, Wzb, and Wzc are translocation proteins that specifically export group I or IV CPS to the outer surface. The identification of wza, wzb, and wzc genes suggested a type of group 1 or IV CPS in K. gyiorum SWMUKG01. It is reported that the initiating glycosyltransferase is the distinguishing factor between group I and IV CPS biosynthesis pathways, with WbaP catalyzing group I and WecA for group IV CPS biosynthesis. By searching in the SWMUKG01 genome, we could not identify a wbaP homologue, but two wecA (PROKKA_02966 and PROKKA_00500) were identified though it locates outside the CPS gene cluster. We proposed that the strain SWMUKG01 appears to have a group IV CPS. As group IV CPS biosynthesis is Wzy-dependent, the O-antigen polymerase Wzy appears to participate in both CPS and LPS biosynthesis in K. gyiorum SWMUKG01, asin Vibrio vulnificus [42].
Analysis on the SWMUKG01 genome sequence identified three genomic islands (GIs) (S5 Table), which were defined by obviously different GC contents in comparison to the average of the genome and associated with the presence of insertion sequences, integrases and transposases. We found that the CPS gene cluster is mainly composed of a specific genomic island GI-III (PROKKA_01918–01931) and several genes on both sides. The GI-III is a 16, 772 bp island with a GC content of 49% and codes for 14 proteins, including those for the biosynthesis and transport of capsular polysaccharides in addition to an IS2 transposase TnpB. The identification of GI-III suggested that these capsular polysaccharide encoding genes might have evolved from a different organism by horizontal gene transfer. Besides, the GI-III presents some characteristics of pathogenicity islands for the presence of some putative virulence related genes (PROKKA_01918, 01919, 01920, 01929 and 01931).
Secretion systems.
The SWMUKG01 genome harbors complete sets of genes coding for proteins that constitute the general secretion (Sec) and two-arginine (Tat) pathways (Table 5), which are essential for export of proteins into the periplasm (Gram-negative bacteria) or plasma membrane (Gram-positive bacteria) [43]. Three sets of closely linked genes encoding the putative T1SS, namely prsD1-prsE1-tolC1 (PROKKA_00929–00931), tolC2-prsD2-prsE2 (PROKKA_01512–01514) and prsE3-prsD3-tolC3 (PROKKA_02054–02056), were identified on this genome (Table 5). The identities among these three T1SSs were 33.5 ± 4.9%, suggesting a structural independence between each other. A further blast showed that prsD1-prsE1-tolC1 of the strain SWMUKG01 had 83% identity and 93% coverage with that from Comamonas Kerstersii, and prsE3-prsD3-tolC3 obtained 70% identity and 92% coverage with that from Variovorax Boronicumulans, but tolC2-prsD2-prsE2 had no significant homologues in the GenBank.
K. gyiorum harbors a putative gsp operon encoding proteins putatively involved in the biosynthesis of T2SS (Table 5). This gsp operon, consisting of gspGHIJKLMDEF, is located between PROKKA_02423 and PROKKA_02433, with an unexpected tRNA encoding gene (PROKKA_02430) inserted into the operon. It seems that GspC and GspO proteins were lacking in the strain SWMUKG01, for that a typical T2SS apparatus bears 12 core components (T2SS CDEFGHIJKLMO) [44]. These gene absences do not necessarily mean that this bacterium lacks a functioning system, because some T2SS proteins, especially C, could have been missed as a result of them being the least conserved among core constituents [44]. More work is needed to determine whether the T2SS homologs in this strain are expressed correctly and encode a functional secretion system.
An interesting finding on the SWMUKG01 genome was the presence of two sets of genes encoding proteins homologous to the conjugation paradigm VirB/D T4SS, which we designated the vir1 (PROKKA_00412–00423) and vir2 (PROKKA_02490–02502) locus (Table 5). The nucleotide sequence identity between these two sets of genes was 77%, indicating that they were very likely to be duplicated copies. Compared with that in vir2 locus, almost the same gene context was contained in vir1, except that double virB4 were found in the former but a single in the latter (Fig 6). Both vir1 and vir2 showed highest identity (75% and 73%) with those of A. aquatilis, and other homologous T4SS were also found in Bordetella petrii, Bordetella trematum, B. bronchiseptica and P. aeruginosa etc (Fig 6). It was unexpected to identify a copG gene between virB11 and virD4 in both vir loci of SWMUKG01. Similar gene arrangement occurs in a conserved fashion in many other bacteria. It is known that CopG is a transcriptional repressor that control the plasmid copy number [45]. And, we found a homologous T4SS encoded by a plasmid pTTS12 in Pseudomonas putida, also with a copG gene between virB11 and virD4. Thus, we proposed that the T4SS encoding system on the SWMUKG01 chromosome is most likely to originate from a plasmid by horizontal transfer. Besides, in comparison with the VirB/VirD4 system of Agrobacterium tumefaciens [46], an archetypal T4SS that is encoded by the vir gene cluster composed of 12 components, virB1–virB11 and virD4, SWMUKG01 lacks virB7 in both vir loci. However, on the SWMUKG01 genome, a hypothetical protein (PROKKA_00419) is placed between virB5 and virB6. Similar gene arrangements also occur in other homologous T4SSs. This suggested that the hypothetical protein may play similar roles as the VirB7 does. If this is true and both vir1 and vir2 are functional, it would be interesting to learn these two T4SS systems coexist in SWMUKG01. Furthermore, the presence of secretion systems in CG1 was investigated; the results showed that all the secretion systems discussed above could be detected except the vir1 locus of T4SS.
The box arrows represent ORFs. Similar regions are indicated with the degree of nucleotide identity being shown in gray scales.
Iron-uptake systems.
Three broad categories of iron-uptake systems have been currently identified: systems for the utilization of ferric iron (Fe3+), ferrous iron (Fe2+) and heme-bound iron. By searching the genome, we found that the SWMUKG01 harbors genes encoding proteins for all three categories of iron-uptake systems.
Approximately 88 genes (2.5%) of the SWMUKG01 genome are involved in iron uptake (Table 6), all of which, except the genes from PROKKA_00721 to PROKKA_00723, could also be detected in CG1. Among them, 16 genes encoding potential outer membrane receptors for ferric siderophores, of which seven are involved in binding ferric enterobactin, four binding ferric dicitrate, two binding ferric-pseudobactin, one binding ferripyoverdine, one binding ferrichrome and one binding ferrichrysobactin. In addition, 8 CDSs for putative TonB-dependent receptor precursors were also identified. Interestingly, all of them are located in a potential operon with fecI and fecR, regulatory genes of the iron dicitrate transport systems, which may indicate a functional relevance between the putative receptor precursors and FecIR for iron uptake. Although the SWMUKG01 possesses multiple outer membrane receptors, each of which provides the bacterium with specificity for different siderophores, it only contains three binding-protein-dependent ABC systems: FepCBDG (PROKKA_00824–00827) for the transport of ferric catechols, FecBCDE (PROKKA_02160–02163) for ferric citrate and FhuBCD (PROKKA_02518–02521) for ferric hydroxamates. In addition, a putative fbp operon (PROKKA_01345–01347) encoding the FbpABC system was found on SWMUKG01 genome. It was previously shown that FbpABC is likely to be a ferric iron transporter that is involved in the translocation of iron delivered by transferrin and lactoferrin, across the cytosolic membrane [47]. Interestingly, no gene encoding putative transferrin or lactoferrin receptor has been found in this organism, leading to the hypothesis that the FbpABC system functions in the utilization of some siderophores as iron source in SWMUKG01 [48].
In K. gyiorum SWMUKG01, five genes encoding CDSs homologous to heme-binding receptors were identified (Table 6). These receptors include one periplasmic heme-binding protein, one hemopexin utilization protein C and three TonB-dependent hemin receptors on the outer membrane, which are likely to be involved in the import of heme by binding it directly or by recognizing its carrier. In addition, orthologs of hmuVUT were found in SWMUKG01, which encode an ABC-type heme uptake system, comprising a periplasmic heme-binding protein HmuT, a permease HmuU and an ATPase HmuV. It is likely that the HmuVUT system is the main participant in delivering the heme to the cytosol, as it is the only potential heme transport system identified in SWMUKG01.
It is demonstrated that Fe2+ is the dominant form of the element under anaerobic and/or acidic conditions [49]. Only one Fe2+-uptake system, EfeUOB, has been identified in SWMUKG01 (Table 6), which is demonstrated to be involved in the uptake of ferrous iron in several bacteria [47]. The efe operon (PROKKA_00585–00587) in this bacterium encodes a ferrous iron permease EfeU and two periplasmic proteins EfeB and EfeO in sequential order, with a slight difference in gene order comparing with previous studies [50]. In Gram-negative bacteria, the inner-membrane anchored TonB/ExbB/ExbD complex provides the energy required to transport the associated cargo across the outer-membrane [51]. A set of genes encoding the TonB system, tonB-exbB-exbD (PROKKA_01872–01870), were identified in SWMUKG01 genome. In some bacteria, such as V. cholerae [52] and P. aeruginosa [53], there is more than one TonB-ExbB-ExbD system, while in SWMUKG01 only one exists. This indicated that the TonB system might be shared by different iron uptake systems and heme transport pathways of SWMUKG01.
Two-component signal transduction systems (TCSs).
In K. gyiorum SWMUKG01, a total of 23 open reading frames were identified as putative RRs, 19 of which are adjacent to genes encoding probable HKs, forming 21 HK/RR pairs, all of which could be identified in CG1 but the gene PROKKA_00858. These histidine kinase and response regulator proteins could be categorized into five groups (Table 7): ten pairs and two single RRs belong to the OmpR subfamily, four pairs fall into the FixJ subfamily, three pairs and a single RR are grouped in the CitB subfamily, one pair is in the NtrC subfamily and the remaining two RRs are members of the CheY subfamily [54]. These TCSs are potentially implicated in regulating several aspects of key processes, such as osmoregulation (EnvZ/OmpR), chemotaxis (CheA/CheY), nitrogen metabolism (NtrY/NtrX), oxygen sensing (AcrA/AcrB) and perhaps pathogenicity mechanisms (QseC/QseB) [55]. Overall, a large number of genes encoding putative TCSs appear to make K. gyiorum well-equipped to respond to and survive environmental changes during the infection cycle.
Antibiotic resistant genes and multidrug efflux pumps.
Clinical reports showed that some of K. gyiorum isolates were resistant to ciprofloxacin [2, 8], colistin [10], cefepime and ceftazidime [7], which suggested potential drug resistance genes or efflux pumps. K. gyiorum SWMUKG01 showed resistant to ciprofloxacin (>2 μg/ml) and cefuroxime (>16 μg/ml). Our genomic analysis showed that 48 (1.39%) out of 3441 potential CDSs were identified in CARD database (S6 Table). Among them, potential resistance genes encoding proteins against fluoroquinolone (gyrB, gyrA), sulfonamide, rifampicin, and fosfomycin were included, as well as some other multidrug resistance proteins. However, no plasmid was detected in the strain SWMUKG01, thus plasmid-carried drug resistance genes are not part of drug resistance for this pathogen.
By searching in the SWMUKG01 genome, we found a total of 10 sets of genes encoding multidrug efflux pump systems (Table 8), among which only two genes PROKKA_00804 and PROKKA_01650 were discovered to be truncated. In these pumps, RND (resistance-nodulation-division) family transporters are most commonly found in SWMUKG01, such as, AcrAB-OprM, MexAB-OprM and BepEF-TtgF. MFS (major facilitator superfamily) (e.g., EmrAB-OprM) and ABC (ATP-binding cassette) (e.g., MacAB-TtgC) efflux pumps are also included. These efflux pumps were reported to be responsible for the direct extrusion of many kinds of drugs, such as aminoglycosides, β-lactams, fluoroquinolones, macrolides and chloramphenicol from the cell [56]. Given the presence of various efflux pumps in SWMUKG01, the risk of efflux-mediated multidrug resistance is a real possibility in this organism.
Conclusions
In conclusion, the complete genome of K. gyiorum strain SWMUKG01, the first clinical isolate from southwest China, was sequenced in our present study. The length of the genome is about 3.9 million bps with genomic GC content of 62%. Genomic and phylogenetic comparisons indicated that K. gyiorum, A. aquatilis and A. faecalis may derive from a recent common ancestor. A total of 3441 CDSs were annotated, of which 326 potential virulence factors were predicted by VFDB database. Genes and operons related to bacterial surface polysaccharides, flagella, pili, iron acquisition systems, secretion systems, and TCSs as well as efflux pumps were analyzed at the genomic level and compared with those from other pathogens, which underlined the genetic basis of the pathogenesis and virulence of K. gyiorum. This work allows the identification of a new bacterial species at the genetic level and provides a foundation for future research into the mechanisms of pathogenesis of K. gyiorum.
Supporting information
S1 Fig. Functional categorization of the K. gyiorum SWMUKG01 genome based on COG database.
https://doi.org/10.1371/journal.pone.0214686.s001
(TIF)
S1 Table. Functional categorization of the K. gyiorum SWMUKG01 genome based on COG database.
https://doi.org/10.1371/journal.pone.0214686.s002
(XLSX)
S2 Table. Classification of genes of K. gyiorum SWMUKG01 by COG codes.
https://doi.org/10.1371/journal.pone.0214686.s003
(XLSX)
S3 Table. Genes encoding proteins with a putative role in virulence of strain SWMUKG01.
https://doi.org/10.1371/journal.pone.0214686.s004
(XLSX)
S4 Table. Genes encoding proteins with a putative role in flagella biosynthesis of strain SWMUKG01.
https://doi.org/10.1371/journal.pone.0214686.s005
(XLSX)
S5 Table. Putative genomic islands in strain SWMUKG01.
https://doi.org/10.1371/journal.pone.0214686.s006
(XLSX)
S6 Table. Putative antibiotic resistant genes in strain SWMUKG01.
https://doi.org/10.1371/journal.pone.0214686.s007
(XLSX)
References
- 1. Coenye T, Vancanneyt M, Cnockaert MC, Falsen E, Swings J, Vandamme P. Kerstersia gyiorum gen. nov., sp. nov., a novel Alcaligenes faecalis-like organism isolated from human clinical samples, and reclassification of Alcaligenes denitrificans Ruger and Tan 1983 as Achromobacter denitrificans comb. nov. Int J Syst Evol Microbiol. 2003;53(Pt 6):1825–31. pmid:14657111.
- 2. Pence MA, Sharon J, McElvania Tekippe E, Pakalniskis BL, Ford BA, Burnham CA. Two cases of Kerstersia gyiorum isolated from sites of chronic infection. J Clin Microbiol. 2013;51(6):2001–4. pmid:23596236; PubMed Central PMCID: PMCPMC3716053.
- 3. Deutscher M, Severing J, Balada-Llasat JM. Kerstersia gyiorum Isolated from a Bronchoalveolar Lavage in a Patient with a Chronic Tracheostomy. Case Rep Infect Dis. 2014;2014:479581. pmid:25506444; PubMed Central PMCID: PMCPMC4259082.
- 4. Vandamme P, De Brandt E, Houf K, De Baere T. Kerstersia similis sp. nov., isolated from human clinical samples. Int J Syst Evol Microbiol. 2012;62(Pt 9):2156–9. pmid:22058318.
- 5. Almuzara MN, Barberis CM, Traglia GM, Ordonez AM, Famiglietti AM, Ramirez MS, et al. Isolation of Kerstersia gyiorum from a patient with cholesteatomatous chronic otitis media. J Clin Microbiol. 2012;50(11):3809–11. pmid:22972825; PubMed Central PMCID: PMCPMC3486247.
- 6. Mwalutende A, Mshana SE, Mirambo MM, Mushi MF, Chalya PL, Giliyoma JM, et al. Two cases of chronic suppurative otitis media caused by Kerstersia gyiorum in Tanzania: is it an underappreciated pathogen in chronic otitis media? Int J Infect Dis. 2014;29:251–3. pmid:25461666.
- 7. Bostwick AD, Zhang C, Manninen K, Touchberry J, Greene SR, Holland TL. Bacteremia Caused by Kerstersia gyiorum. J Clin Microbiol. 2015;53(6):1965–7. pmid:25809974; PubMed Central PMCID: PMCPMC4432044.
- 8. Uysal EB, Celik C, Tuzcu N, Can F, Dogan M, Erturk R, et al. A case of chronic suppurative otitis media caused by Kerstersia gyiorum. APMIS. 2015;123(11):986–9. pmid:26303793.
- 9. Ogawa Y, Lee ST, Kasahara K, Koizumi A, Chihara Y, Nakano R, et al. A first case of isolation of Kerstersia gyiorum from urinary tract. J Infect Chemother. 2016;22(4):265–7. pmid:26683244.
- 10. Baran I, Duzgun AP, Mumcuoglu I, Aksu N. Chronic lower extremity wound infection due to Kerstersia gyiorum in a patient with Buerger's disease: a case report. BMC Infect Dis. 2017;17(1):608. pmid:28877672; PubMed Central PMCID: PMCPMC5588736.
- 11. Borsa BA, Kaplan HH, Bayri Baris A, Gungordu Dalar Z, Ozalp VC. Kerstersia gyiorum: An Unusual Pathogen Causing Chronic Suppurative Otitis Media. Klimik Dergisi/Klimik Journal. 2017;30(3):158–60.
- 12. Lan Y, Yan Q, Yan Y, Liu W. First case of Kerstersia gyiorum isolated from a patient with chronic osteomyelitis in China. Frontiers in Laboratory Medicine. 2017;1(3):141–3.
- 13. Kim JH, Lee E, Lee Y. The First Case of Chronic Otitis Media due to Kerstersia gyiorum in Korea. Ann Lab Med. 2018;38(6):607–9. pmid:30027707; PubMed Central PMCID: PMCPMC6056389.
- 14. Mardis ER. The impact of next-generation sequencing technology on genetics. Trends Genet. 2008;24(3):133–41. pmid:18262675.
- 15. Greninger AL, Kozyreva V, Truong CL, Longoria R, Chaturvedi V. Draft Genome Sequence of Kerstersia gyiorum CG1, Isolated from a Leg Ulcer. Genome Announc. 2015;3(5). pmid:26358603; PubMed Central PMCID: PMCPMC4566185.
- 16. Chin CS, Alexander DH, Marks P, Klammer AA, Drake J, Heiner C, et al. Nonhybrid, finished microbial genome assemblies from long-read SMRT sequencing data. Nat Methods. 2013;10(6):563–9. pmid:23644548.
- 17. Seemann T. Prokka: rapid prokaryotic genome annotation. Bioinformatics. 2014;30(14):2068–9. pmid:24642063.
- 18. Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, Miller W, et al. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 1997;25(17):3389–402. pmid:9254694; PubMed Central PMCID: PMCPMC146917.
- 19. Tatusov RL, Fedorova ND, Jackson JD, Jacobs AR, Kiryutin B, Koonin EV, et al. The COG database: an updated version includes eukaryotes. BMC bioinformatics. 2003;4:41. pmid:12969510; PubMed Central PMCID: PMCPMC222959.
- 20. Kanehisa M, Goto S, Kawashima S, Okuno Y, Hattori M. The KEGG resource for deciphering the genome. Nucleic Acids Res. 2004;32(Database issue):D277–80. pmid:14681412; PubMed Central PMCID: PMCPMC308797.
- 21. Kanehisa M, Goto S, Hattori M, Aoki-Kinoshita KF, Itoh M, Kawashima S, et al. From genomics to chemical genomics: new developments in KEGG. Nucleic Acids Res. 2006;34(Database issue):D354–7. pmid:16381885; PubMed Central PMCID: PMCPMC1347464.
- 22. Magrane M, UniProt C. UniProt Knowledgebase: a hub of integrated protein data. Database (Oxford). 2011;2011:bar009. pmid:21447597; PubMed Central PMCID: PMCPMC3070428.
- 23. Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, Cherry JM, et al. Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nat Genet. 2000;25(1):25–9. pmid:10802651; PubMed Central PMCID: PMCPMC3037419.
- 24. Stothard P, Wishart DS. Circular genome visualization and exploration using CGView. Bioinformatics. 2005;21(4):537–9. pmid:15479716.
- 25. Jia B, Raphenya AR, Alcock B, Waglechner N, Guo P, Tsang KK, et al. CARD 2017: expansion and model-centric curation of the comprehensive antibiotic resistance database. Nucleic Acids Res. 2017;45(D1):D566–D73. pmid:27789705; PubMed Central PMCID: PMCPMC5210516.
- 26. Chen L, Yang J, Yu J, Yao Z, Sun L, Shen Y, et al. VFDB: a reference database for bacterial virulence factors. Nucleic acids research. 2005;33(Database issue):D325–8. pmid:15608208; PubMed Central PMCID: PMCPMC539962.
- 27. Bertelli C, Laird MR, Williams KP, Simon Fraser University Research Computing G, Lau BY, Hoad G, et al. IslandViewer 4: expanded prediction of genomic islands for larger-scale datasets. Nucleic Acids Res. 2017;45(W1):W30–W5. pmid:28472413; PubMed Central PMCID: PMCPMC5570257.
- 28. Price MN, Dehal PS, Arkin AP. FastTree 2—approximately maximum-likelihood trees for large alignments. PLoS One. 2010;5(3):e9490. pmid:20224823; PubMed Central PMCID: PMCPMC2835736.
- 29. Zhao Y, Wu J, Yang J, Sun S, Xiao J, Yu J. PGAP: pan-genomes analysis pipeline. Bioinformatics. 2012;28(3):416–8. pmid:22130594; PubMed Central PMCID: PMCPMC3268234.
- 30. Gao F, Zhang CT. Ori-Finder: a web-based system for finding oriCs in unannotated bacterial genomes. BMC Bioinformatics. 2008;9:79. pmid:18237442; PubMed Central PMCID: PMCPMC2275245.
- 31. Richter M, Rossello-Mora R. Shifting the genomic gold standard for the prokaryotic species definition. Proceedings of the National Academy of Sciences of the United States of America. 2009;106(45):19126–31. pmid:19855009; PubMed Central PMCID: PMCPMC2776425.
- 32. Nejidat A, Saadi I, Ronen Z. Effect of flagella expression on adhesion of Achromobacter piechaudii to chalk surfaces. J Appl Microbiol. 2008;105(6):2009–14. pmid:19120648.
- 33. Lee JJ, Srinivasan S, Kim MK. Pigmentiphaga soli sp. nov., a bacterium isolated from soil. J Microbiol. 2011;49(5):857–61. pmid:22068507.
- 34. Parkhill J, Sebaihia M, Preston A, Murphy LD, Thomson N, Harris DE, et al. Comparative analysis of the genome sequences of Bordetella pertussis, Bordetella parapertussis and Bordetella bronchiseptica. Nat Genet. 2003;35(1):32–40. pmid:12910271.
- 35. Wadhams GH, Armitage JP. Making sense of it all: bacterial chemotaxis. Nat Rev Mol Cell Biol. 2004;5(12):1024–37. pmid:15573139.
- 36. Tomich M, Planet PJ, Figurski DH. The tad locus: postcards from the widespread colonization island. Nat Rev Microbiol. 2007;5(5):363–75. pmid:17435791.
- 37. Kachlany SC, Planet PJ, DeSalle R, Fine DH, Figurski DH. Genes for tight adherence of Actinobacillus actinomycetemcomitans: from plaque to plague to pond scum. Trends Microbiol. 2001;9(9):429–37. pmid:11553455.
- 38. Pelicic V. Type IV pili: e pluribus unum? Mol Microbiol. 2008;68(4):827–37. pmid:18399938.
- 39. Yuan B, Cheng A, Wang M. Polysaccharide export outer membrane proteins in Gram-negative bacteria. Future Microbiol. 2013;8(4):525–35. pmid:23534363.
- 40. Raetz CR, Whitfield C. Lipopolysaccharide endotoxins. Annu Rev Biochem. 2002;71:635–700. pmid:12045108; PubMed Central PMCID: PMCPMC2569852.
- 41. Allison GE, Verma NK. Serotype-converting bacteriophages and O-antigen modification in Shigella flexneri. Trends Microbiol. 2000;8(1):17–23. pmid:10637639.
- 42. Nakhamchik A, Wilde C, Rowe-Magnus DA. Identification of a Wzy polymerase required for group IV capsular polysaccharide and lipopolysaccharide biosynthesis in Vibrio vulnificus. Infect Immun. 2007;75(12):5550–8. pmid:17923517; PubMed Central PMCID: PMCPMC2168344.
- 43. Tseng TT, Tyler BM, Setubal JC. Protein secretion systems in bacterial-host associations, and their description in the Gene Ontology. BMC Microbiol. 2009;9 Suppl 1:S2. pmid:19278550; PubMed Central PMCID: PMCPMC2654662.
- 44. Cianciotto NP. Type II secretion: a protein secretion system for all seasons. Trends Microbiol. 2005;13(12):581–8. pmid:16216510.
- 45. Hernandez-Arriaga AM, Rubio-Lepe TS, Espinosa M, del Solar G. Repressor CopG prevents access of RNA polymerase to promoter and actively dissociates open complexes. Nucleic Acids Res. 2009;37(14):4799–811. pmid:19520770; PubMed Central PMCID: PMCPMC2724298.
- 46. Li PL, Hwang I, Miyagi H, True H, Farrand SK. Essential components of the Ti plasmid trb system, a type IV macromolecular transporter. Journal of bacteriology. 1999;181(16):5033–41. pmid:10438776; PubMed Central PMCID: PMCPMC93993.
- 47. Andrews SC, Robinson AK, Rodríguez-Quiñones F. Bacterial iron homeostasis. FEMS Microbiology Reviews. 2003;27(2–3):215–37. pmid:12829269
- 48. Strange HR, Zola TA, Cornelissen CN. The fbpABC operon is required for Ton-independent utilization of xenosiderophores by Neisseria gonorrhoeae strain FA19. Infect Immun. 2011;79(1):267–78. pmid:21041493; PubMed Central PMCID: PMCPMC3019921.
- 49. Wei Y, Murphy ER. Shigella Iron Acquisition Systems and their Regulation. Front Cell Infect Microbiol. 2016;6:18. pmid:26904516; PubMed Central PMCID: PMCPMC4746246.
- 50. Cao J, Woodhall MR, Alvarez J, Cartron ML, Andrews SC. EfeUOB (YcdNOB) is a tripartite, acid-induced and CpxAR-regulated, low-pH Fe2+ transporter that is cryptic in Escherichia coli K-12 but functional in E. coli O157:H7. Mol Microbiol. 2007;65(4):857–75. pmid:17627767.
- 51. Miethke M, Marahiel MA. Siderophore-based iron acquisition and pathogen control. Microbiol Mol Biol Rev. 2007;71(3):413–51. pmid:17804665; PubMed Central PMCID: PMCPMC2168645.
- 52. Mey AR, Payne SM. Haem utilization in Vibrio cholerae involves multiple TonB-dependent haem receptors. Mol Microbiol. 2001;42(3):835–49. pmid:11722746.
- 53. Zhao Q, Poole K. A second tonB gene in Pseudomonas aeruginosa is linked to the exbB and exbD genes. FEMS microbiology letters. 2000;184(1):127–32. pmid:10689178.
- 54. Gao R, Tao Y, Stock AM. System-level mapping of Escherichia coli response regulator dimerization with FRET hybrids. Molecular Microbiology. 2008;69(6):1358–72. pmid:18631241
- 55. Gotoh Y, Eguchi Y, Watanabe T, Okamoto S, Doi A, Utsumi R. Two-component signal transduction as potential drug targets in pathogenic bacteria. Curr Opin Microbiol. 2010;13(2):232–9. pmid:20138000.
- 56. Du D, Wang-Kan X, Neuberger A, van Veen HW, Pos KM, Piddock LJV, et al. Multidrug efflux pumps: structure, function and regulation. Nat Rev Microbiol. 2018;16(9):523–39. pmid:30002505.