Antibiotic resistance has emerged globally as one of the biggest threats to human and animal health. Although the excessive use of antibiotics is recognized as accelerating the selection for resistance, there is a growing body of evidence suggesting that natural environments are "hot spots" for the development of both ancient and contemporary resistance mechanisms. Given that pharmaceuticals can be entrained onto agricultural land through anthropogenic activities, this could be a potential driver for the emergence and dissemination of resistance in soil bacteria. Using functional metagenomics, we interrogated the "resistome" of bacterial communities found in a collection of Canadian agricultural soil, some of which had been receiving antibiotics widely used in human medicine (macrolides) or food animal production (sulfamethazine, chlortetracycline, and tylosin) for up to 16 years. Of the 34 new antibiotic resistance genes (ARGs) recovered, the majority were predicted to encode (multi)drug efflux systems, while a few share little to no homology with established resistance determinants. We characterized several novel gene products, including putative enzymes that can confer high-level resistance against aminoglycosides, sulfonamides, and broad range of beta-lactams, with respect to their resistance mechanisms and clinical significance. By coupling high-resolution proteomics analysis with functional metagenomics, we discovered an unusual peptide, PPP AZI 4 , encoded within an alternative open reading frame not predicted by bioinformatics tools. Expression of the proline-rich PPP AZI 4 can promote resistance against different macrolides but not other ribosome-targeting antibiotics, implicating a new macrolidespecific resistance mechanism that could be fundamentally linked to the evolutionary design of this peptide.IMPORTANCE Antibiotic resistance is a clinical phenomenon with an evolutionary link to the microbial pangenome. Genes and protogenes encoding specialized and potential resistance mechanisms are abundant in natural environments, but understanding of their identity and genomic context remains limited. Our discovery of several previously unknown antibiotic resistance genes from uncultured soil microorganisms indicates that soil is a significant reservoir of resistance determinants, which, once acquired and "repurposed" by pathogenic bacteria, can have serious impacts on therapeutic outcomes. This study provides valuable insights into the diversity and identity of resistance within the soil microbiome. The finding of a novel peptide-mediated resistance mechanism involving an unpredicted gene product also highlights the usefulness of integrating proteomics analysis into metagenomicsdriven gene discovery.