Alu interspersed elements are inserted into the genome by a retroposition process that occurs via dimeric Alu RNA and causes genetic disorders in humans. Alu RNA is labile and can be diverted to a stable left monomer transcript known as small cytoplasmic Alu (scAlu) RNA by RNA 3 processing, although the relationship between Alu RNA stability, scAlu RNA production, and retroposition has been unknown. In vivo, Alu and scAlu transcripts interact with the Alu RNA-binding subunit of signal recognition particle (SRP) known as SRP9/14. We examined RNAs corresponding to Alu sequences that were differentially active during primate evolution, as well as an Alu RNA sequence that is currently active in humans. Mutations that accompanied Alu RNA evolution led to changes in a conserved structural motif also found in SRP RNAs that are associated with thermodynamic destabilization and decreased affinity of the Alu right monomer for SRP9/14. In contrast to the right monomer, the Alu left monomer maintained structural integrity and high affinity for SRP9/14, indicating that scAlu RNA has been under selection during human evolution. Loss of Alu right monomer affinity for SRP9/14 is associated with scAlu RNA production from Alu elements in vivo. Moreover, the loss in affinity coincided with decreased rates of Alu amplification during primate evolution. This indicates that stability of the Alu right monomer is a critical determinant of Alu retroposition. These results provide insight into Alu mobility and evolution and into how retroposons may interact with host proteins during genome evolution.At nearly 1 million copies in primate DNA, Alu interspersed elements are the most successful transposons known. The great majority of Alu repeats in human DNA were fixed in an ancestral primate genome before the emergence of the human lineage (reviewed in reference 40). Although Alu retroposition indeed occurs in humans (16,35,46,48), the available data indicate that (i) certain Alu sequences have been more prolific than others and (ii) the rate of new Alu insertions into the genome has declined during recent periods of primate evolution (6,40,41).The Alu retroposons that were actively proliferating during ancient, intermediate, and modern evolutionary times are reflected by three subfamilies of Alu sequences that remain distinguishable in human DNA (7,23,38,44,50; reviewed in references 40 and 41). The subfamily consensus sequence referred to as Alu Sx (formerly PS and Major; see nomenclature in reference 2) represents the sequence that produced approximately 85% of the Alus in the human genome 60 to 30 million years ago. The Alu Y sequence (formerly CS and Precise) was active 30 to 15 million years ago and represents ϳ15% of Alu sequences in human DNA. The Alu Ya5 consensus (formerly HS and PV) represents the sequence that was active over the last 5 million years that produced Ͻ0.5% of the Alu sequences in human DNA (2,3,14,22,27,33,38,40,41). These data argue that the average Alu retroposition rate in the human lineage has been only 1% tha...