backtranseq http://emboss.sourceforge.net/apps/release/6.3/emboss/apps/backtranseq.html

backtranseq reads a protein sequence and writes the nucleic acid sequence it is most likely to have come from.

This is not very efficient, but it does work (notice, these files do not have headers)..

cat file.aa
MFLILLI

cat file.map
M       AAA
F       TTT
L       CCC
I       GGG

for x in $(fold -w 1 file.aa); do awk -v x="$x" 'BEGIN{FS="\t"}{if($1==x){printf $2}}' file.map; done
AAATTTCCCGGGCCCCCCGGG

But surely you are aware that it's impossible to go from aa sequence back to the original nt sequence? E.g. six different codons encode Leucine..