Greetings all,
Here is my general idea... not all that elegant I know but it does what the OP is asking... I think.

#!/usr/bin/perl -w
use strict;
my $incoming_xml = '
<?xml version="1.0" encoding="UTF-8"?>
<TRANSACTION>
    <FIELDS>
        <FIELD KEY="user">name</FIELD>
        <FIELD KEY="password">pass&word</FIELD>
        <FIELD KEY="operation_type"><do_what></FIELD>
     </FIELDS>
</TRANSACTION>';
print $incoming_xml;
print "\n==============\n";
$incoming_xml =~ s/(<FIELD[^>]*>)(.*)(<\/FIELD>)/$1.&html_transliterat
+e($2).$3/eg;
print $incoming_xml;
exit;
sub html_transliterate{
    my $in_str = shift;
    $in_str =~ s/&/!38/g;
    $in_str =~ s/</!40/g;
    $in_str =~ s/>/!41/g;
    return $in_str;
}
[download]

Output

<?xml version="1.0" encoding="UTF-8"?>
<TRANSACTION>
    <FIELDS>
        <FIELD KEY="user">name</FIELD>
        <FIELD KEY="password">pass&word</FIELD>
        <FIELD KEY="operation_type"><do_what></FIELD>
     </FIELDS>
</TRANSACTION>
==============

<?xml version="1.0" encoding="UTF-8"?>
<TRANSACTION>
    <FIELDS>
        <FIELD KEY="user">name</FIELD>
        <FIELD KEY="password">pass!38word</FIELD>
        <FIELD KEY="operation_type">!40do_what!41</FIELD>
     </FIELDS>
</TRANSACTION>
[download]

of course this is only working for XML nodes <FIELD> right now but can easily be modified. HTH.