Discovering potent antibiotics is of critical importance due to the substantial increases of microbial resistance. Xanthones are intriguing sources of antimicrobials, despite a scarcity of extensive investigations into their mechanisms of action. Here, we reported the development of a series of xanthone derivatives, among which compound XT17 displayed strong broad-spectrum antibacterial activity, weak hemolytic activity and low cytotoxicity against mammalian cell lines, low frequencies of drug resistance, and potent in vivo efficacy in Staphylococcu aureus or Pseudomonas aeruginosa induced murine corneal infection models. Compound XT17 presented a multifaceted mode of actions, involving the disruption of cell wall by interacting with lipoteichoic acid or lipopolysaccharides and the suppression of DNA synthesis. A further docking study confirmed the capability of compound XT17 to form a stable complex with the bacterial gyrase enzyme. This work could offer an innovative design strategy for developing broad-spectrum therapeutic agents against drug-resistant bacteria.

A series of xanthone derived antimicrobial agents were designed and synthesized. The xanthone scaffold serves as the core precursor for a further chemically modification at its phenolic hydroxyl positions to improve the pharmacological properties of xanthone derivatives against both Gram-positive and Gram-negative bacteria. The candidate compound XT17, identified through a structure-activity relationship (SAR) study of these synthesized compounds, showed strong bactericidal properties and potent in vitro efficacy. A comprehensive transcriptomic analysis of the candidate compound XT17 was assessed using the high-throughput RNA-seq technique to gain insight into the antibacterial mechanisms. This efficacy stems from a dual-action mechanism: cell wall destruction and DNA gyrase inhibition.

Growth condition and total RNA extraction

For the challenge experiments, 5 ml of E. coli ATCC 25922 at an OD₆₀₀ of 0.5, representing the mid-log phase, was exposed to compound XT17 (4× MIC) for 2 h in a biological duplicate. Cultures treated with 4× MIC ciprofloxacin and without antibiotics served as positive and negative controls, respectively. After exposure, 1 ml aliquots of cells were immediately pelleted at 4 °C by centrifugation for 2 min at 2000 rpm. The supernatants were removed and immediately frozen in liquid nitrogen. The samples were stored at −80 °C for later total RNA isolation. Total RNA was extracted using the bacterial RNA kit (Omega Bio-Tek, USA) according to the manufacturer’s instructions. The RNA Integrity Number (RIN) values were obtained to assess RNA quality using an Agilent Genomics 2200 Tape Station instrument.

Transcriptome sequencing

A total amount of 2–5 ng of the RNA per sample was executed as the input material to construct each cDNA library for RNA sequencing using the NEBNext Ultra Directional RNA Library prep kit from Illumina. The quality of the resulting libraries was checked using Agilent High Sensitivity DNA chips to ensure proper library size distribution and the absence of small adapters. Libraries were quantified and normalized by qPCR before being sequenced at 150 cycles with the NextSeq 500 High Output Kit, yielding approximately 9 million, 75 bp, and paired-end reads for each library.

Bioinformatics analysis

Read mapping was aligned using Bowtie2. HTSeq v0.6.1 was used for quantification of expression levels. DESeq2 on the R package (v1.12.4) was used to calculate differentially expressed genes. Differential expression was computed using edgeR: exactTest for each treatment versus the untreated control. The E. coli database used for mapping was GCF_017357505.1_ASM1735750v1_genomic.fna. Both the Gene Ontology (GO), and Kyoto Encyclopedia of Genes and Genomes (KEGG Ortholog database) enrichment analyses were done with Fisher’s exact test. All P values were adjusted with the Benjamini-Hochberg procedure to generate false-discovery rates (FDRs; adjusted P values). GO enrichment analysis was carried out using the clusterProfiler package on the R platform. KEGG pathway mapping and figure generation were performed with the PathView website. Clusters of Orthologous Groups (COG) of protein databases with an E-value of 1 e^-5. For hierarchal clustering analysis of gene expression, the gene expression levels were normalized to z-scores of log₁₀ (FPKM + 1), where FPKM is fragments per kilobase per million. For hierarchal clustering analysis of the regulation of transcription factors, the UMPGA method was used as the clustering method. Transcription of genes that were not notably regulated was considered unchanged. Volcano plots were generated by plot_volcano from soothsayer (https://github.com/jolespin/soothsayer) in Python v.3.6.6. Directed networks (DiNetwork) were constructed and plotted using NetworkX and Matplotlib Python packages, respectively. Hierarchical clustering analysis and heatmap calculation were performed with the pheatmap package on R platform. The umap package on the R platform was used for UMAP analysis of FPKM vectors of each transcriptome. The figures were created via ggplot2 package.