The host environment activates multiple stress responses which correlate with antibiotic tolerance in Staphylococcus aureus

Abstract

Introduction. Antibiotic tolerance allows genetically susceptible bacteria to survive antibiotic exposure and contributes to treatment failure and recurrent infection. In Staphylococcus aureus, tolerance is closely associated with biofilm infections, and this association is conventionally explained by the biofilm mode of growth itself, in which mass transfer limitation is presumed to starve and arrest cells in the deeper layers. Gap Statement. That explanation rests largely on biofilms grown in rich laboratory media, and it has rarely been tested against the alternative that the host environment, rather than the biofilm architecture, drives the underlying physiological change. Furthermore, because transcriptomic and bulk fluorescence methods average across the population, they cannot establish whether stress responses are activated uniformly in all cells or strongly in a minority. Aim. To determine, at single-cell resolution, whether bacterial stress responses are activated by biofilm growth, by the host environment, or by both, and whether such activation coincides with a reduced rate of antibiotic killing. Methodology. Promoter regions of recA, katA, relQ, ilvleu and groESL were fused to a stable green fluorescent protein variant to report the SOS, oxidative stress, cell wall stress- and starvation-activated stringent, and heat shock responses. Reporter activation was quantified by confocal microscopy in exponential and stationary cultures, after defined chemical and physical stresses, in biofilms grown in tryptic soy broth, in diluted and undiluted human serum, in heat-inactivated serum, and in S. aureus internalised by human neutrophils. Tolerance was assessed as the rate of killing in time-kill assays. Results. Biofilms grown in laboratory media or in diluted serum showed minimal stress response activation, whereas biofilms of the same strain grown in undiluted human serum activated all five responses, uniformly across the population. Heat inactivation of the serum did not abolish this effect, indicating that complement and heat-labile serum enzymes are not required. One hour of serum exposure activated the stringent response via both starvation and cell wall stress, and it abolished detectable killing by ciprofloxacin, while producing only small changes in killing by vancomycin and rifampicin and none by daptomycin or dicloxacillin. Pre-treatments that activated the stringent or heat shock responses likewise arrested growth and slowed ciprofloxacin killing. Neutrophil phagocytosis activated the oxidative stress response and both stringent response branches. The host environment, rather than the biofilm mode of growth, is the dominant driver of stress response activation in S. aureus. Activation is population-wide and coincides with population-wide ciprofloxacin tolerance. Laboratory biofilm models that omit host components are therefore likely to underestimate the tolerance encountered during infection.