Resistance to colistin is increasingly reported in Klebsiella pneumoniae clinical isolates. The aim of this study was to analyze the molecular epidemiology and virulence profiles of 25 colistin-resistant K. pneumoniae blood isolates from the Hospital Agency “Ospedale dei Colli,” Naples, Italy, during 2015 and 2016. Colistin MIC values of isolates ranged from 4 to 256 mg/L. The inactivation of the mgrB gene, encoding a negative regulator of the PhoQ/PhoP signaling system, was the most frequent mechanism of colistin resistance found in 22 out of 25 isolates. Of these, 10 isolates assigned to ST512 and PFGE types A and A4 showed identical frameshift mutation and premature termination of mgrB gene; 4 isolates assigned to ST258 and PFGE types A1 showed non-sense, frameshift mutation, and premature termination; 3 and 1 isolates assigned to ST258 and PFGE A2 and ST512 and PFGE A3, respectively, had insertional inactivation of mgrB gene due to IS5-like mobile element; 2 isolates assigned to ST101 and 1 to ST392 had missense mutations in the mgrB gene, 1 isolate assigned to ST45 showed insertional inactivation of mgrB gene due to IS903-like mobile element. phoQ missense mutations were found in 2 isolates assigned to ST629 and ST101, respectively, which also showed a missense mutation in pmrA gene. The mcr-1-2-3-4 genes were not detected in any isolate. Colistin-resistant K. pneumoniae isolates showed variable virulence profiles in Galleria mellonella infection assays, with the infectivity of two isolates assigned to ST45 and ST629 being significantly higher than that of all other strains (P < 0.001). Interestingly, colistin MIC values proved to make a significant contribution at predicting lethal doses values (LD50 and LD90) of studied isolates in G. mellonella. Our data show that MgrB inactivation is a common mechanism of colistin resistance among K. pneumoniae in our clinical setting. The presence of identical mutations/insertions in isolates of the same ST and PFGE profile suggests the occurrence of clonal expansion and cross-transmission. Although virulence profiles differ among isolates irrespective of their genotypes, our results suggest that high colistin MIC could predict lower infectivity capability of the isolates.

Modifications of the lipid A moiety of lipopolysaccharide influence the physicochemical properties of the outer membrane of Gram-negative bacteria. Some bacteria produce lipid A with a single hydroxylated secondary acyl chain. This hydroxylation is catalyzed by the dioxygenase LpxO, and is important for resistance to cationic antimicrobial peptides (e.g., polymyxins), survival in human blood, and pathogenicity in animal models. The lipid A of the human pathogen Pseudomonas aeruginosa can be hydroxylated in both secondary acyl chains, but the genetic basis and physiological role of these hydroxylations are still unknown. Through the generation of single and double deletion mutants in the lpxO1 and lpxO2 homologs of P. aeruginosa PAO1 and lipid A analysis by mass spectrometry, we demonstrate that both LpxO1 and LpxO2 are responsible for lipid A hydroxylation, likely acting on different secondary acyl chains. Lipid A hydroxylation does not appear to affect in vitro growth, cell wall stability, and resistance to human blood or antibiotics in P. aeruginosa. In contrast, it is required for infectivity in the Galleria mellonella infection model, without relevantly affecting in vivo persistence. Overall, these findings suggest a role for lipid A hydroxylation in P. aeruginosa virulence that could not be directly related to outer membrane integrity.

No abstract

Colistin represents a last-line treatment option for infections caused by multidrug resistant Gram-negative pathogens, including Pseudomonas aeruginosa. Colistin resistance generally involves the modification of the lipid A moiety of lipopolysaccharide (LPS) with positively charged molecules, namely phosphoethanolamine (PEtN) or 4-amino-4-deoxy-L-arabinose (Ara4N), that reduce colistin affinity for its target. Several lines of evidence highlighted lipid A aminoarabinosylation as the primary colistin resistance mechanism in P. aeruginosa, while the contribution of phosphoethanolamination remains elusive. PEtN modification can be due to either endogenous (chromosomally encoded) PEtN transferase(s) (e.g., EptA in P. aeruginosa) or plasmid borne MCR enzymes, commonly found in enterobacteria. By individually cloning eptA and mcr-1 into a plasmid for inducible gene expression, we demonstrated that MCR-1 and EptA have comparable PEtN transferase activity in P. aeruginosa and confer colistin resistance levels similar to those provided by lipid A aminoarabinosylation. Notably, EptA, but not MCR-1, negatively affects P. aeruginosa growth and, to a lesser extent, cell envelope integrity when expressed at high levels. Mutagenesis experiments revealed that PEtN transferase activity does not account for the noxious effects of EptA overexpression, that instead requires a C-terminal tail unique to P. aeruginosa EptA, whose function remains unknown. Overall, this study shows that both endogenous and exogenous PEtN transferases can promote colistin resistance in P. aeruginosa, and that PEtN and MCR-1 mediated resistance has no impact on growth and cell envelope homeostasis, suggesting that there may be no fitness barriers to the spread of mcr-1 in P. aeruginosa.

No abstract