Dynamics of Antibacterial Drone Establishment in Staphylococcus aureus: Unexpected Effects of Antibiotic Resistance Genes

The antibacterial drone (ABD) system is based on repurposing the phage-inducible staphylococcal pathogenicity islands (SaPIs) for use as antibacterial agents that are indifferent to antibiotic resistance. The ABDs were constructed by inserting tetM for tetracycline resistance (Tcr) selection, replacing the SaPI virulence genes with bactericidal or bacteriostatic genes such as CRISPR/cas9/agrA, whose expression kills by double-strand cleavage of agrA, or CRISPR/dcas9/agrP(2)P(3), whose expression blocks the target organism's virulence. ABD DNA is packaged in phage-like particles that attack their staphylococcal targets in vivo as well as in vitro. We determine ABD titers by transfer frequency, enumerate surviving cells as a function of multiplicity, and analyze the fate of ABD DNA with green fluorescent protein. An initial study revealed surprisingly that many more cells were killed by the ABD than were measured by transduction. Our study of this phenomenon has revealed several important features of the ABD system: (i) a significant number of entering ABD DNA molecules do not go on to establish stable transductants (i.e., are abortive); (ii) ABD cargo genes are expressed immediately following entry, even by the abortive ABDs; (iii) immediate plating on Tc-containing agar seriously underestimates particle numbers, partly owing to Tc inhibition of protein synthesis; (iv) replacement of tetM with cadA (conferring resistance to CdCl2) provides more accurate particle enumeration; (v) ABDs expressing CRISPR/cas9/agrA kill similar to 99.99% of infected cells and provide the most accurate measurement of particle numbers as well as proof of principle for the system; and (vi) surprisingly, TetM interferes with stable establishment of ABD DNA independently of Tc-r. IMPORTANCE For a particulate therapeutic agent, such as the ABD, accurate enumeration of particles is critical to enable evaluation of preparative procedures and calculation of therapeutic dosages. It is equally important that a selective marker used for these two purposes be biologically inert. We have long used tetM for these purposes but show here that tetM not only underestimates particle titers, by over 20-fold in some experiments, but also seriously impedes stable establishment of the therapeutic particle DNA. Given that tetM is a very convenient and widely used selective marker, publication of these findings is of considerable importance to the microbiological community as well as an interesting illustration of the unpredictable biological effects of genes taken out of their native context.