Host adaptation and convergent evolution increases antibiotic resistance without loss of virulence in a major human pathogen

Abstract

As human population density and antibiotic exposure increases, specialised bacterial subtypes have begun to emerge. Arising among species that are common commensals and infrequent pathogens, antibiotic-resistant high-risk clones have evolved to better survive in the modern human. Here, we show that the major matrix porin (OmpK35) of Klebsiella pneumoniae is not required in the mammalian host for colonisation, pathogenesis, nor for antibiotic resistance, and that it is commonly absent in pathogenic isolates. This is found in association with, but apparently independent of, a highly specific change in the co-regulated partner porin, the osmoporin (OmpK36), which provides enhanced antibiotic resistance without significant loss of fitness in the mammalian host. These features are common in well-described high-risk clones of K. pneumoniae, as well as in unrelated members of this species and similar adaptations are found in other members of the Enterobacteriaceae that share this lifestyle. Available sequence data indicates evolutionary convergence, with implications for the spread of lethal antibiotic-resistant pathogens in humans.\n\nAuthor summaryKlebsiella pneumoniae is a Gram-negative enterobacteria and a significant cause of human disease. It is a frequent agent of pneumonia, and systemic infections can have high mortality rates (60%). OmpK35 and OmpK36 are the major co-regulated outer membrane porins of K. pneumoniae. OmpK36 absence has been related to antibiotic resistance but decreased bacterial fitness and diminished virulence. A mutation that constricts the porin channel (Gly134Asp135 duplication in loop 3 of the porin, OmpK36GD) has been previously observed and suggested as a solution to the fitness cost imposed by loss of OmpK36.\n\nIn the present study we constructed isogenic mutants to verify this and test the impact of these porin changes on antimicrobial resistance, fitness and virulence. Our results show that loss of OmpK35 has no significant cost in bacterial survival in nutrient-rich environments nor in the mammalian host, consistent with a predicted role outside that niche. When directly compared with the complete loss of the partner osmoporin OmpK36, we found that isogenic OmpK36GD strains maintain high levels of antibiotic resistance and that the GD duplication significantly reduces neither gut colonisation nor pathogenicity in a pneumonia mouse model. These changes are widespread in unrelated genomes. Our data provide clear evidences that specific variations in the loop 3 of OmpK36 and the absence of OmpK35 in K. pneumoniae clinical isolates are examples of successful adaptation to human colonization/infection and antibiotic pressure, and are features of a fundamental evolutionary shift in this important human pathogen.