Photocrosslinking allows first-order structural analysis with relatively small amounts of biological material and can be applied in complex in vitro systems. In this article we describe methods for positioning both arylazide and thionucleotide photoagents within an RNA of interest by end modification of circularly permuted RNAs. Application of this technique provided a library of constraints that, together with biochemical and phylogenetic comparative data, were used to develop a structure model of the bacterial ribonuclease P ribozyme-substrate complex. Circularly permuted genes for in vitro transcription are generated by PCR from tandem genes. Circularly permuted RNA transcripts can be modified with high efficiency at both the 5' and 3' termini with arylazide crosslinking reagents, or transcription can be primed with photoactive nucleotide analog monophosphates such as 6-thioguanosine. These crosslinking agents can be used over a wide range of experimental conditions but remain inert until they are activated by UV light. Crosslinked sites are subsequently mapped by reverse transcriptase primer extension of gel-purified crosslinked species. In addition to providing basic protocols for these methods, we discuss approaches for establishing the relevance of crosslinking data to native RNA structure.