Potential in-host evolution of Klebsiella pneumoniae ST147: convergence and the role of capsular alterations in morphotype diversity

Abstract

Klebsiella pneumoniae, an important opportunistic pathogen, is traditionally classified into classic and hypervirulent pathotypes. Convergent strains combining antimicrobial resistance (AMR) and hypervirulence have emerged globally, posing a significant health challenge. In this study, we investigated potential in-host evolution and morphotypic variation among consecutive K. pneumoniae ST147 isolates from a single patient. As described in a previous study, the isolates displayed distinct morphotypes-small white, normal white, gray, and gray/dry (g/d) colonies. In this study, we now present new genomic and phenotypic analyses, which suggest the acquisition of AMR and virulence genes through plasmid gain. The early isolates carried fewer plasmids, resulting in low resistance and virulence, while later isolates acquired a hybrid IncFIB-IncHI1B plasmid encoding different resistance and aerobactin genes, significantly increasing their geno- and phenotypic AMR and virulence potential. Chromosomal integration of this plasmid in one isolate stabilized the acquired traits. Disruptions in the K locus, mediated by insertion sequences, were linked to the gray and g/d morphotypes, impairing capsule biosynthesis. Uronic acid assays confirmed reduced capsule production in these isolates. In contrast, small colony variants showed significant transcriptomic changes, including upregulation of capsule biosynthesis, iron uptake pathways, and AMR genes, suggesting persistence through altered metabolism. Our findings suggest in-host microevolution of K. pneumoniae ST147 from a classic to a convergent pathotype and highlight the genomic and transcriptomic adaptations underlying morphotypic diversity, providing new insights into the pathogen's adaptability and persistence in certain environments.IMPORTANCEKlebsiella pneumoniae is an important bacterial pathogen that can cause severe infections, particularly in healthcare settings. Traditionally, strains are classified as either drug-resistant or highly virulent, but recent strains combining both features have led to hard-to-treat infections. In this study, we investigated the potential in-host evolution of K. pneumoniae ST147 isolates from a single patient, revealing key genomic and phenotypic changes driving their adaptability. Over time, isolates acquired additional plasmids carrying antimicrobial resistance and virulence genes, including a hybrid plasmid integrated into the bacterial chromosome, stabilizing its beneficial traits. In addition, morphotypic diversity emerged, linked to genomic disruptions in capsule biosynthesis genes (K locus). While capsule-deficient morphotypes displayed structural changes, small colony variants exhibited transcriptomic adaptations to persist under stress. This work underscores the dynamic evolutionary capacity of K. pneumoniae adapting within the host, providing critical insights into its persistence, resilience, and the challenges posed by emerging convergent strains in clinical settings.

Keywords: K. pneumoniae; SCV; capsule deficiencies; convergent pathotype; in-host evolution.

Fig 1

Temporal dynamics of plasmid carriage and virulence-associated features in white phenotype isolates. (A) Schematic representation of plasmid carriage in the isolates. Plasmid carriage in isolates 2A to 5B introduced additional virulence genes characteristic of hypervirulent K. pneumoniae, including the aerobactin translocation system (iuc, iut) and the mucoid phenotype regulator (rmpA2). Notably, the rmpA2 gene was only partially present, with approximately 40% coverage. (B) Quantification of capsular polysaccharides by uronic acid assay (n = 3). Data are expressed as mean ± standard error of the mean. (C) Siderophore secretion levels expressed as mean percentage units of siderophore production ± standard error of the mean (n = 3). (D) Kaplan-Meier survival analysis of G. mellonella larvae after infection (n = 30 larvae/condition). Results are expressed as mean percent mortality after injection of 10⁵ CFU per larva. Statistical comparisons for uronic acid and siderophore secretion assays were conducted using one-way ANOVA with Dunnett’s post hoc test against isolate 1A, and for the survival analysis, log-rank tests were performed against isolate 1A; significance levels are indicated as ***, P < 0.001; ****, P < 0.0001. PC: positive control.

Fig 2

Chromosomal integration of a multireplicon IncFIB-IncHI1B plasmid in isolate 5D. (A) Schematic representation of the plasmid content and chromosomal integration event in isolate 5D compared to isolate 5B. (B) Synteny plot comparing the chromosome of isolate 5B, the chromosome of isolate 5D, and the IncFIB-IncHI1B plasmid. The alignment shows the integration of the IncFIB-IncHI1B plasmid region (highlighted in teal) into the 5D chromosome. The position of the custom primer pairs used to demonstrate chromosomal plasmid integration is indicated by arrows above (1F and 1R) and below (2F and 2R) the chromosome of 5D. Selected resistance and virulence genes, such as bla_CTX-M-15, bla_NDM-1, iutA, and iucA, are annotated for clarity. The corresponding regions have high sequence similarity (99%–100%), as indicated by the gradient scale, and show the extensive genomic rearrangements associated with plasmid integration. For simplicity, the iucBCD operon has been omitted from the annotations.

Fig 3

Overview of K and O locus interruptions and deletions. The multiple sequence alignment shows the K (KL64) and O (O1/O2v1) loci of all isolates in comparison to the public reference sequence of K. pneumoniae strain W281 (CP162992). Filled black rectangles in the reference show the reference backbone, while horizontal black lines indicate gaps. Gray rectangles for the isolates indicate sequences identical with the reference, while horizontal black lines indicate gaps, and vertical lines indicate SNP positions (see legend). Black rectangles for the isolates indicate insertions (IS1R).

Fig 4

UpSet plot representing shared and uniquely DEGs in isolates 5A-D. All phenotypes (5A: small; 5C: gray; 5D: g/d) were compared to the white morphotype (5B), and only significantly different genes are shown (log2FC ≥ 2). Bars indicate the number of significantly up- or downregulated genes per isolate. Overlapping responses (intersections) are defined below the bars with connected dots. As the UpSet plot breaks down the DEGs into overlaps between isolates, they need to be summed accordingly to obtain the full count per isolate.

Fig 5

COG analysis for DEG in isolate 5A.