The sites of transcription initiation or RNA processing are generally determined by S1 protection or primer extension (PE) analysis. Although S1 protection must be used for 3'-end analysis, 5'-end analysis is usually more easily & precisely done by primer extension analysis. In this method, a short antisense 5' end-labeled DNA primer (usually a synthetic oligonucleotide, but sometimes a small restriction fragment) is hybridized to RNA, usually total cellular RNA, then DNA is synthesized from this primer using reverse transcriptase. RTase will copy the RNA from the site of primer annealing to the 5'-end of the RNA molecule. The reactions are then analyzed by electrophoresis in sequencing gels in lanes adjacent to Sanger sequencing reactions of DNA containing the gene of question, using the same primer as that used in the PE analysis. The transcription initiation site (usually) can then easily be identified as the band in the sequencing reaction directly parallel to the run-off reverse transcript. Multiple transcription initiation sites will appear as multiple bands in the PE lane. RNA processing sites (i.e. in the case of stable RNAs) will also appear in PE analyses, and must be distinguished from transcription initiation sites by other means (usually by examination of the DNA sequence for promoter elements).
Before you begin, of course, you must have the sequence of the region in question and a DNA primer in a region that must be in the transcript, i.e. the upstream part of the ORF or structural RNA. Usually, the primers are 18-25 base synthetic oligonucleotides. The method described is unsuitable for eukaryotic RNAs, unless it is known that no introns (which would give misleading results) are present between the 5' end of the RNA & the primer. You must also pre-make sequencing ladders, for standards, of the region being analyzed. These are easily prepared from cloned DNA using the same primer used in the PE analysis and either sequenase or klenow chain termination sequencing procedures. Make enough for several gels, or , if you prefer, prpare the sequencing ladders fresh for each PE experiment (this is probably best, if you know the sequencing works well each time). Lastly, you will need RNA. Total cellular RNA is OK - it should be undegraded and purifed free of DNA, for example by the CsCl/EtBr method. You'll need alot of RNA - 125ug per lane on the gel, or 250ug per experiment. The RNA is probably the most important ingredient of the experiment - good RNA, good results. Remember, use RNase-free technique whenever handling RNA!
Transcription & Translation (Hames & Higgins, eds) IRL Press
JWBrown, unpublished data
GSBeckler, PhD thesis