Transposons, or “jumping genes,” are segments of DNA that autonomously mobilize within genomes. During transposition, self-encoded transposases synapse the left and right transposon ends, packaging the mobile element for integration. These ends can be appended to exogenous DNA, making transposons attractive genome-editing tools. These ends can also be asymmetrical, providing control over transposon insertion orientation. However, end sequences can be hundreds of base-pairs long, limiting available target sites and cargo design. It remains unclear why multiple transposase binding sites are present on ends when one transposase subunit catalyzes strand transfer per end. Further, how a single protein recognizes an asymmetric arrangement of DNA-binding sites has eluded mechanistic characterization. Here, we present the cryo-EM structure of a transposase in complex with transposon DNA ends. An associated DNA-bending host factor is critical for full occupancy of transposon binding sites. Extensive DNA flexibility resulted in conformational heterogeneity. To obtain an interpretable map, we performed multiple local map refinements to generate a high-resolution composite. This strategy produced an atomic model highlighting the unique architecture of the asymmetric assembly, providing insight into transposon self-recognition. We found that transposases bound to non-terminal binding sites form inter-end pairing interfaces, revealing a functional role for subunits not directly catalyzing strand-transfer. This pairing requires precise DNA distortions coordinated by both the transposase and DNA-bending host factor. Our structure shows how asymmetrical ends create an architecture incompatible with symmetric ends, enabling left-right end distinction. Sparse protein-DNA contacts suggest that transposases may not require strict sequence specificity for end recognition. Instead, we propose that DNA distortions induced by the transposase and DNA-bending proteins cooperatively create “allowable” configurations compatible with end synapsis. Together, these findings elucidate multiple factors contributing to transposon end recognition, providing a foundation for redesign to expand genome-editing applications.
Truong et al. (Sun,) studied this question.