Mu transposition occurs within a large protein-DNA complex called a transpososome. This stable complex includes four subunits of MuA transposase, each contacting a 22-base pair recognition site located near an end of the transposon DNA. These MuA recognition sites are critical for assembling the transpososome. Here we report that when concentrations of Mu DNA are limited, the MuA recognition sites permit assembly of transpososomes in which non-Mu DNA substitutes for some of the Mu sequences. These "hybrid" transpososomes are stable to competitor DNA, actively transpose the non-Mu DNA, and produce transposition products that had been previously observed but not explained. The strongest activator of non-Mu transposition is a DNA fragment containing two MuA recognition sites and no cleavage site, but a shorter fragment with just one recognition site is sufficient. Based on our results, we propose that MuA recognition sites drive assembly of functional transpososomes in two complementary ways. Multiple recognition sites help physically position MuA subunits in the transpososome plus each individual site allosterically activates transposase.Transposons are found in all the biological kingdoms, and some perform specialized functions. For example, the machinery that initiates V(D)J recombination likely evolved from a transposon (1, 2), and the cDNA of HIV and other retroviruses integrate into host cell DNA through mechanisms nearly identical to transposition (3). The genome of bacteriophage Mu is a transposon that uses transposition both to integrate into the DNA of a new host cell and to replicate before lysis. Like most DNA rearrangements, transposition is a complex, multi-step process, requiring numerous DNA sequence elements. Studies of bacteriophage Mu have been central to our understanding of both the fundamental mechanisms and the complexities of DNA transposition.Phage Mu encodes a transposase, MuA, that transfers the Mu genome from one DNA location (the transposition donor) to a new location (the transposition target) (4, 5). During transposition, transposase performs two principle reactions: DNA cleavage and DNA strand transfer. During cleavage, the donor DNA is nicked twice, once at each 3Ј end of the Mu genome. During strand transfer, the cleaved transposon ends are inserted into neighboring sites on the two target strands.Little or no specific sequence information is needed on the target DNA (6), but the Mu DNA provides many sequence cues for transposition (Fig. 1). For example, the last two nucleotides at either 3Ј end of the Mu DNA, the cleavage sites, have the sequence 5Ј-CA. Also near each end of the Mu DNA are three recognition sites, distinct from the cleavage sites, which share a 22-base pair consensus sequence. The recognition sites are referred to as R1, R2, and R3 on the right end and L1, L2, and L3 on the left end ( Fig.