Fig. 1: Drug–drug interactions are species-specific in Gram-positive bacteria.
From: Systematic analysis of drug combinations against Gram-positive bacteria
a, Schematic representation of the high-throughput screen. Pairwise combinations of 65 drugs belonging to several chemical classes and targeting different cellular processes (Supplementary Table 2) were tested at three concentrations in S. aureus (two strains), B. subtilis and S. pneumoniae. For each strain, 1,891–2,070 combinations were tested in broth microdilution in 384-well plates, measuring OD595 over time. Normalized fitness values were calculated and used to obtain 4 × 4 checkerboards and assign interactions as synergistic, antagonistic or neutral (Methods, Extended Data Fig. 1 and Supplementary Table 3). PMF, proton-motive force. b, Interaction abundance in each strain separately and altogether. Synergy and antagonism frequencies were obtained by dividing their absolute counts by the number of combinations for which the probed fitness space allows detection of synergy (fitness upon combination ≥0.1) or antagonism (fitness upon combination ≤0.9) discovery (Methods). Total numbers of combinations tested (n) and detected interactions (i) are shown for each set. c, Conservation of interactions among the four strains tested. All unique interactions detected in the screen (n = 725) were considered to calculate intersection sets between strains. The total number of interactions dependent on whether conserved or unique to each strain/species are shown. A total of 81 interactions (i), involving 47 drugs (d), are conserved across species (dark red). The total number of interactions in each strain is indicated as set size (bottom right), adding up to 945 total interactions in all strains. d, Network of conserved interactions between Gram-positive species. Drugs are grouped according to their targeted cellular process (Supplementary Table 2). Edge thickness is proportional to the number of drug–drug interactions for each class–class pair. Node size is proportional to the number of drugs in each class. Only drugs involved in this interaction set are considered (d = 47). Nodes are coloured according to the targeted cellular processes as in Fig. 1a. e, Drug interaction conservation between species recapitulates phylogeny. Pearson correlation between sequence identity (based on 40 conserved marker genes) and drug interaction conservation rate is between pairs of species tested here and previously11. The P value was obtained from a two-tailed one-sample t-test assessing the significance of the Pearson correlation (H0: {t = 0, R = 0}).
