A slightly different approach:

c:\@Work\Perl\monks>perl -wMstrict -le
"my $Dna1 = qq{AACAGCACGGCAACGCTGTGCCTTGGGCACCATGCAGTACCAAACGGAACGATAG
+TGAAAACAATCACGA\n};
 ;;
 while ($Dna1 =~ /C[AG]G/g) {
   my ($start, $end, $len) = ($-[0], $+[0], $+[0]-$-[0]);
   print qq{pattern C[AG]G matched @ offsets $start thru $end, $len lo
+ng};
   }
"
pattern C[AG]G matched @ offsets 2 thru 5, 3 long
pattern C[AG]G matched @ offsets 7 thru 10, 3 long
pattern C[AG]G matched @ offsets 34 thru 37, 3 long
pattern C[AG]G matched @ offsets 44 thru 47, 3 long
[download]

c:\@Work\Perl\monks>perl -wMstrict -le
"my $Dna1 = qq{AACAGCACGGCAACGCTGTGCCTTGGGCACCATGCAGTACCAAACGGAACGATAG
+TGAAAACAATCACGA\n};
 ;;
 while ($Dna1 =~ /(C[AG]G)/g) {
   my $sub_seq = $1;
   my ($start, $end, $len) = ($-[1], $+[1], $+[1]-$-[1]);
   print qq{matched '$sub_seq' @ offsets $start thru $end, $len long};
   }
"
matched 'CAG' @ offsets 2 thru 5, 3 long
matched 'CGG' @ offsets 7 thru 10, 3 long
matched 'CAG' @ offsets 34 thru 37, 3 long
matched 'CGG' @ offsets 44 thru 47, 3 long
[download]

Update: Reading the OP more closely, I see you want to capture and print the matched sub-sequence also. I have posted new code to do this using a capture group. Sorry for any confusion. (Update: Buried the evidence a bit deeper under some  <readmore> tags.)