In this study, we investigated interactions between temperature and antibiotics at sub-MIC concentrations for P. fluorescens and showed that this interaction was twofold: at low and very high temperatures (possibly outside the organisms’ temperature optimum), the antibiotic mixtures showed increased synergy, yet overall temperature increased antibiotic efficacy for single antibiotics. A striking result of our study was that antibiotic mixtures had lethal effects even when the concentrations added together were below their respective individual toxicity in a realistic antibiotic pollution scenario.

The latter result was unexpected because ciprofloxacin and ofloxacin both belong to the fluoroquinolone class and are not expected to act in synergy. Antibiotics with different target action are more likely to show interactions when combined—either as synergy [20] or as antagonism/suppression [10]—and there is now growing evidence that both synergy and antagonism are a common feature of antibiotics, and generally pharmaceuticals and other stressors, in mixtures [10,18,21]. Focusing on antibiotic mixing is important as risk assessments and ecotoxicological tests are based on single compounds but antibiotics, in common with all pharmaceuticals, do not occur as isolated and pure substances in the environment and they should be regarded as a multi-component chemical mixture [2]. A growing body of literature shows that mixtures of pollutants can have different effects compared to single compounds and that the joint effect of such chemical cocktails is often higher than the toxicity of each individual compound [12,22,23]. For example, González-Pleiter et al. (2013) demonstrated that the combined effect can be ‘more than the sum of the parts’ by testing the effects of antibiotics in mixtures including a mixture of erythromycin and tetracycline that had particular strong synergistic effects on cyanobacteria [11]. However, knowledge about the toxicity of antibiotic mixtures is still limited and ignoring possible mixture effects might underestimate the actual impact of antibiotics in the environment [20].

We observed low- and high-temperature-dependent synergistic effects in inhibiting bacterial growth, suggesting antibiotic-specific off-target effects that only affect the bacteria once they enter a certain temperature range. This is in line with emerging literature on antibiotic mixing and temperature [5,14,23] that highlights the context dependency of antibiotic efficacy and it alludes to the fact that stressors such as chemicals, temperature or pH interact [5]—especially when levels are reached that are outside or ‘on the edges’ of the organism’s tolerance breadth [14,24]. Complicating matters, these interactions ‘play out’ on different levels of biological organization—from subcellular to the individual and population level. For example, while it is intuitive that low concentrations of antibiotics change populations because they can provoke resistance in bacteria [8], even more complex mechanisms are at play here and antibiotic stress below lethal levels can result in bacterial strains with narrowed temperature breadth and shifted temperature optima [24]—resulting in individuals/populations that are more susceptible to stressors.

In our study, EC₅₀ values of single antibiotics decreased with temperature and explanations for this pattern include that both uptake [25] and metabolism [16] of antibiotics increase with temperature. Additionally, high temperatures enhance the toxicity of contaminants (yet, at the same time, enhance the rates of chemical degradation [26]). Further, synergistic effects are possible such as both temperature and (some) antibiotics influencing protein folding and synthesis [27,28]. Cruz-Loya and colleagues (2019) found that cellular responses to temperature stress have likely been evolutionarily co-opted to also respond to many classes of antibiotic stress [14]. A further factor is that a population-level effect could come into play in a nutrient-limited environment, as a rise in temperature results in increased population density and potentially competition for resources. In our experiments, the assays with higher population density resulted in lower EC₅₀ values of ciprofloxacin and ofloxacin compared to the experiment with lower bacterial densities. Density-dependent effects in bacteria could therefore be explored more when it comes to antibiotic assays and studies, as is indeed the case for other driving forces of evolution in bacteria such as time, space or disturbance (but see, e.g., [29]).

Interactions among different stressors [21] are at the core of unexpected ecological impact because interactions can lessen or amplify the direct signal effect of each stressor [30]. In this vein, adaptation to both temperature and antibiotics is another future research avenue and a strong focus is needed on sub-lethal antibiotic concentrations as highlighted above. Changes in environmental temperature ‘hit’ multicellular organisms in ‘acute’ ways (such as species extinctions or range shifts [31]) but also shape microbial communities despite their seemingly immediate capability to adapt. For instance, bacterial strains adapted to high temperatures can be more sensitive to certain antibiotics [14] and generally temperature can alter the average body size of microbes (e.g., [32]) and this in turn affects metabolic rates [16,33,34].