I'd recommend writing a short wrapper around e.g. MAFFT or similar and then script your way out of keeping track of the structure encoding. Once the sequences are aligned, it should be simple to repeat the '-' chars comparing each aligned sequence to its unaligned counterpart...

Thank you very much  for your response. It is not a problem to write the script I just wanted to ensure that there isn't a widely used program which does that automatically...

Aligning tRNA sequences by an algorithm based software??  Whoops! Its not so straightforward like Proteins......errors could be incorporated especially based on some irregular presence or absence of some nucleotides in selected positions like 16, 17, 20a, 20b, 47 etc. Also sequences of tRNA obtained from Databases are sometimes not accurate.......especially in selecting the start/end positions. I hope this helps....! 

Yes, I know. I'm looking for RNAs using a script combining  RNA scanner and Aragorn and I'm planning to align it using models in Infernal. I think it should be OK....