Klebsiella pneumoniae Asparagine tDNAs Are Integration Hotspots for Different Genomic Islands Encoding Microcin E492 Production Determinants and Other Putative Virulence Factors Present in Hypervirulent Strains

Abstract

Due to the developing of multi-resistant and invasive hypervirulent strains, Klebsiella pneumoniae has become one of the most urgent bacterial pathogen threats in the last years. Genomic comparison of a growing number of sequenced isolates has allowed the identification of putative virulence factors, proposed to be acquirable mainly through horizontal gene transfer. In particular, those related with synthesizing the antibacterial peptide microcin E492 (MccE492) and salmochelin siderophores were found to be highly prevalent among hypervirulent strains. The determinants for the production of both molecules were first reported as part of a 13-kbp segment of K. pneumoniae RYC492 chromosome, and were cloned and characterized in E. coli. However, the genomic context of this segment in K. pneumoniae remained uncharacterized. In this work, we provided experimental and bioinformatics evidence indicating that the MccE492 cluster is part of a highly conserved 23-kbp genomic island (GI) named GIE492, that was integrated in a specific asparagine-tRNA gene (asn-tDNA) and was found in a high proportion of isolates from liver abscesses sampled around the world. This element resulted to be unstable and its excision frequency increased after treating bacteria with mitomycin C and upon the overexpression of the island-encoded integrase. Besides the MccE492 genetic cluster, it invariably included an integrase-coding gene, at least seven protein-coding genes of unknown function, and a putative transfer origin that possibly allows this GI to be mobilized through conjugation. In addition, we analyzed the asn-tDNA loci of all the available K. pneumoniae assembled chromosomes to evaluate them as GI-integration sites. Remarkably, 73% of the strains harbored at least one GI integrated in one of the four asn-tDNA present in this species, confirming them as integration hotspots. Each of these tDNAs was occupied with different frequencies, although they were 100% identical. Also, we identified a total of 47 asn-tDNA-associated GIs that were classified into 12 groups of homology differing in theencoded functionalities but sharing with GIE492 a conserved recombination module and potentially its mobility features. Most of these GIs encoded factors with proven or potential role in pathogenesis, constituting a major reservoir of virulence factors in this species.

Keywords: asparagine tRNA gene; hypervirulent Klebsiella pneumoniae; liver abscess; microcin E492; pathogenicity island; salmochelin.

FIGURE 1

The MccE492-production gene cluster is located inside a 23-kbp genomic region with biased GC content and codon usage. A 500-kbp region of the KpRYC492 chromosome (coordinates 1,466,268–1,966,267) was analyzed (n: distance in base-pairs to the start of the segment). GC profile algorithm calculations (A) indicated that two segmentation points (blue arrows) delimitate a 23-kbp region (dark yellow area) comprising MccE492 production cluster. This region showed an average GC content of 44% (B). Codon adaptation index (CAI) calculations revealed that genes located inside this region also show a strong bias in codon usage (C). The red dot corresponds to the CAI value for the integrase gene located between the MccE492 coding gene and the asn-tRNA gene. Gray arrows in (A) indicate two additional segmentation points delimitating a further region unrelated to GIE492 that could also be acquired horizontally.

FIGURE 2

GIE492 features and genomic context. GIE492 is a 22,291-bp DNA segment flanked by direct repeats (DR1 and DR2). It harbors an integrase-coding gene, the previously characterized MccE492-production gene cluster, and at least seven additional protein-coding genes of unknown function, provisionally designed u1 to u7. This GI is inserted in one of the four copies of the asn-tDNA present in K. pneumoniae (asn1A to asn1D). These asn1 loci cluster together in a 20-kbp domain with a gene organization conserved in all the strains that do not harbor asn1-GIs inserted, as the case of the chromosome of K. pneumoniae MGH78578 reference strain. In this scheme, we show the asn1 domain of MGH78578 and particularly asn1C, where GIE492 is integrated in the equivalent region of KpRYC492 chromosome.

FIGURE 3

Detection of GIE492 excision. (A) Schematic representation of the excision event. Direct repeats are depicted as blue boxes. Two primer sets (P1/P2, purple arrows; and P3/P4, blue arrows) were designed to hybridize the regions outside and next to the island borders. Before excision, P3 and P4 delimitate a region of more than 22 kbp. After excision, these primers become 246-pb apart from each other, delimitating a region (scar) susceptible to be amplified by PCR. (B) PCR-amplification of the scar using genomic DNA of KpRYC492 extracted in exponential (E) or stationary (S) phase of growth and primers P1 to P4. A first amplification with P1+P2 and a subsequent amplification with P3+P4 (nested PCR) were required to detect the scar, visualized as a prominent band of the expected size (black arrow). (C) Partial nucleotide sequence of the bands obtained in (B) confirming that they correspond to the scar left after GIE492 excision, including the 17-bp perfect repeat (shaded in light blue).

FIGURE 4

GIE492 excision frequency increased upon addition of mitomycin C to the culture medium or overexpressing the island-encoded integrase. Excision frequency was measured by qPCR starting from genomic DNA isolated from KpRYC492 cells cultured until different phases of growth in a medium without or supplemented with mitomycin C (mito C). (A) Growth curves of KpRYC492 in LB medium supplemented with up to 2 μg/ml mito C. Arrows indicate the points from which genomic DNA was extracted. (B) Dose-dependent mito C-mediated increase of GIE492 excision frequency. (C) Effect of overexpressing the island-encoded integrase over the excision frequency. KpRYC492 cells were transformed with either pint (allowing IPTG-inducible expression of int gene) or pCA24N (control) and grown in LB supplemented with IPTG, or IPTG plus mito-C until the indicated phase of growth. Error bars correspond to the standard deviation of two measurements from three independent experiments. ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001.

FIGURE 5

Genes from GIE492 encoded in the MccE492-production cluster and genes unrelated to this bacteriocin production are transcribed at comparable levels. Relative abundances of GIE492-encoded transcripts were measured by qRT-PCR, starting from total RNA isolated from KpRYC492 cells cultured until mid-exponential or late-exponential phase of growth. Expression values were calculated considering the amplification efficiencies of each primer pair used to amplify every target gene. rpoD gene was used to normalize. Error bars indicate the standard deviation among four measurements from two independent experiments.

FIGURE 6

asn-tDNA loci as integration hotspots for genomic islands in Klebsiella pneumoniae. (A) A total of 47 GIs (colored squares) integrated in any of the four asn-tRNA loci (asn1A to asn1D) were identified among the assembled chromosomes of 52 K. pneumoniae strains. Those GIs can be classified into 12 homology groups (see color code). (B) Schematic representation of the main features of each asn1-GI group. Blue rectangles and red arrows represent direct repeats and integrase-coding genes, respectively. Black and green arrowheads represent insertion sequences and transfer origins (oriT), respectively. ¹Previously described as GI-I (Lery et al., 2014); ²Previously described as ICEKp1 (Lin et al., 2008). R–M, restriction–modification; HPs, hypothetical proteins of unknown function; MT, methyltransferase; NA, nucleic acids.