Off-rates of proteins from the DNA double helix are widely considered to be dependent only on the interactions inside the initially bound protein-DNA complex and not on the concentration of nearby molecules. However, a number of recent single-DNA experiments have shown off-rates that depend on solution protein concentration, or "facilitated dissociation." Here, we demonstrate that this effect occurs for the major Escherichia coli nucleoid protein Fis on isolated bacterial chromosomes. We isolated E. coli nucleoids and showed that dissociation of green fluorescent protein (GFP)-Fis is controlled by solution Fis concentration and exhibits an "exchange" rate constant (k exch ) of Ϸ10 4 M ؊1 s ؊1 , comparable to the rate observed in single-DNA experiments. We also show that this effect is strongly salt dependent. Our results establish that facilitated dissociation can be observed in vitro on chromosomes assembled in vivo. A ll aspects of chromosome dynamics involve the binding and unbinding of proteins from the DNA double helix. The rate of binding for a given species of protein to a location along a DNA is usually controlled by concentration of that species, but it is widely assumed that unbinding times, or equivalently off-rates, are controlled by the strength of interactions inside the DNA-protein complex and are independent of other nearby molecules in the nucleoplasm (1). However, this picture has been challenged by a number of recent in vitro experiments which have shown off-rates of proteins from duplex DNA that depend on bulk protein concentrations (2-5). Similar effects have been observed for singlestranded DNA (ssDNA)-binding proteins (6, 7), antibody-ligand binding (8), and ssDNA-ssDNA interactions (9).A straightforward explanation for this effect is that macroscopic dissociation (complete dissociation followed by diffusive motion of a protein to a location far away from the initial binding site) occurs via one or more partially dissociated intermediate ("microdissociated") states. A protein in bulk solution could impact dissociation events in a variety of ways (5, 9). This interaction can affect the rate of macroscopic dissociation of the original protein, for example, by having the second "invading" protein "block" rebinding of the first protein (5, 10, 11). Alternatively, both the initially bound and invading proteins might bind at adjacent sites, with the second protein accelerating the off-rate of the first protein via allosteric interactions (12). In either case, formation of a transient ternary complex (initially bound protein plus DNA site plus invading protein) may lead to appreciable increases in off-rates as the bulk protein concentration is increased. In the molecularly crowded cell interior, the rates of replacement of proteins on DNA may well be very different from what might be predicted from dilute-solution studies of binding kinetics, a fact that may help reconcile observations of rapid turnover on DNA in vivo but slower kinetics observed in vitro (13)(14)(15)(16)(17).An example o...