One resistance gene on its own rarely makes a pathogen untreatable — a single defense answers a single drug. Multidrug resistance (MDR) appears when several defenses pile up inside the same cell, and there are only a few recognizable ways that happens.
The first route is physical linkage. Resistance genes sitting on the same plasmid, transposon, or integron are inherited as a unit, so one acquisition event delivers several defenses at once. An integron is a genetic element that captures resistance gene cassettes and expresses them from a shared promoter; a single integron can carry cassettes for several unrelated drug classes. Because the genes are linked, selection for any one of them preserves the whole array.
The second route needs no new genes. A multidrug efflux pump such as AcrAB-TolC in Escherichia coli spans the inner and outer membrane and physically expels structurally unrelated compounds, so one regulatory mutation that switches the pump into overexpression raises the minimum inhibitory concentration for several classes at once. Here resistance is not a collection of separate genes but one physiological change with a wide effect.
The third route is slower and cumulative. A strain that already carries one plasmid can take up a second element, such as a transposon, and recombination between mobile elements can stitch the determinants into new combinations. Each round of antibiotic exposure filters the population toward cells carrying more defenses, because those cells survive more treatment regimens.
These routes are not mutually exclusive. A clinical isolate may carry a beta-lactamase on one plasmid, an aminoglycoside-modifying enzyme on a transposon, and an overexpressed efflux pump — three different molecular logics producing resistance to three classes in one cell. The practical consequence is that resistance to one drug is a poor predictor of susceptibility to another, so treating an MDR infection means testing several agents rather than assuming that an untried drug will work.