First, don't use non-ASCII characters in your code unless you absolutely must have them there (e.g. as a value to be assigned to a variable). As pointed out above, using non-ASCII "smart quotes" to delimit literal strings in your code is an error.

Next, if you want to use $1 ($2, etc) after doing a regex match, you have to put parens into the regex to capture some part(s) of what is being matched.

Apart from those two points, there's not much that needs to be added to your code:

#!/usr/bin/perl
use warnings;
use strict;
use diagnostics;
my $Dna1 = "AACAGCACGGCAACGCTGTGCCTTGGGCACCATGCAGTACCAAACGGAACGATAGTGA
+AAACAATCACGA\n";
          
while ($Dna1 =~ /(C[AG]G)/g) {
      my $endposition = pos($Dna1) + 1;
      print "Pattern C[AG]G matched $1 ending at $endposition\n";
}
[download]

When I run that, I get:

Pattern C[AG]G matched CAG ending at 6
Pattern C[AG]G matched CGG ending at 11
Pattern C[AG]G matched CAG ending at 38
Pattern C[AG]G matched CGG ending at 48
[download]

Those offset values (6, 11, 38, 48) represent the position of the next character after the 3-letter match (where the first character of the string is at position 1). That is, "6" points to the "C" that follows the first "CAG", "11" points to the "C" that follows the next "CGG", and so on.

(updated to fix a typo)